Issue
=====
While listing datacenters associated with a zone, only zone Id validation is required.
There is no need to have additional checks like zone is a legacy zone or not.
Fix
===
Removed unnecessary checks over zone ID and just checking if zone with specified ID exists or not.
Signed-off-by: Sateesh Chodapuneedi <sateesh.chodapuneedi@accelerite.com>
(cherry picked from commit 0ef1c17541)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Summary: CLOUDSTACK-8921
snapshot_store_ref table should store actual size of back snapshot in secondary storage
Calling SR scan to make sure size is updated correctly
(cherry picked from commit 4e4b67cd96)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
with XenServer & Local SR (Db_exn.Uniqueness_constraint_violation)
removed the host uuid from SR label so that any host which has access to
the SR(all the hosts in the same pool) can reuse the same SR
(cherry picked from commit 1aa6a72bc7)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
as that snapshot will never be going to use again and also it will fill up primary storage
(cherry picked from commit 336df84f17)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Updated hardcoded value with max data volumes limit from hypervisor capabilities.
(cherry picked from commit 93f5b6e8a3)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Libvirt / Qemu (KVM) does not collect statistics about these either.
On some systems it might even yield a 'internal error' from libvirt
when attempting to gather block statistics from such devices.
For example Ubuntu 16.04 (Xenial) has a issue with this.
Skip them when looping through all devices.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
The 'force' option provided with the stopVirtualMachine API command is
often assumed to be a hard shutdown sent to the hypervisor, when in fact
it is for CloudStacks' internal use. CloudStack should be able to send
the 'hard' power-off request to the hosts.
When forced parameter on the stopVM API is true, power off (hard shutdown)
a VM. This uses initial changes from #1635 to pass the forced parameter
to hypervisor plugin via the StopCommand, and fixes force stop (poweroff)
handling for KVM, VMware and XenServer.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
* 4.9:
Do not set gateway to 0.0.0.0 for windows clients
CLOUDSTACK-9904: Fix log4j to have @AGENTLOG@ replaced
ignore bogus default gateway when a shared network is secondary the default gateway gets overwritten by a bogus one dnsmasq does the right thing and replaces it with its own default which is not good for us so check for '0.0.0.0'
Activate NioTest following changes in CLOUDSTACK-9348 PR #1549
CLOUDSTACK-9828: GetDomRVersionCommand fails to get the correct version as output Fix tries to return the output as a single command, instead of appending output from two commands
CLOUDSTACK-3223 Exception observed while creating CPVM in VMware Setup with DVS
CLOUDSTACK-9787: Fix wrong return value in NetUtils.isNetworkAWithinNetworkB
with XenServer & Local SR (Db_exn.Uniqueness_constraint_violation)
removed the host uuid from SR label so that any host which has access to
the SR(all the hosts in the same pool) can reuse the same SR
1. Removed XenServerGuestOsMemoryMap from CitrixHelper.java
This java file was holding a static in memory map named XenServerGuestOsMemoryMap. This was the source for xenserver dynamic memory values(max and min). These values were moved to guest_os_details table.
2. DAO layer was modified to access these values.
3. VirtualMachineTo object was modified to populate the dynamic memory values.
4. addGuestOs and UpdateGuestOS api has been modified to update memory values.
This fixes log4j xml to have @AGENTLOG@ replaced with values defined
in build/replace.properties.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Test scenarios:
- Enable cluster HA after VR is created. Now stop and start VR and check its restart priority, should be High.
- Enable cluster HA before VR is created. Now create some VM and verify that VR created must have High restart priority.
CLOUDSTACK-9720: [VMware] template_spool_ref table is not getting updated with correct template physical size in template_size column.Updated the template_spool_ref table with the correct template (VMware - OVA file) size.
* pr/1880:
CLOUDSTACK-9720: [VMware] template_spool_ref table is not getting updated with correct template physical size in template_size column.
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
Summary: CLOUDSTACK-8921
snapshot_store_ref table should store actual size of back snapshot in secondary storage
Calling SR scan to make sure size is updated correctly
Bug-ID: CLOUDSTACK-8880: calculate free memory on host before deploying Vm. free memory = total memory - (all vm memory)With memory over-provisioning set to 1, when mgmt server starts VMs in parallel on one host, then the memory allocated on that kvm can be larger than the actual physcial memory of the kvm host.
Fixed by checking free memory on host before starting Vm.
Added test case to check memory usage on Host.
Verified Vm deploy on Host with enough capacity and also without capacity
* pr/847:
Bug-ID: CLOUDSTACK-8880: calculate free memory on host before deploying Vm. free memory = total memory - (all vm memory)
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
CLOUDSTACK-9794: Unable to attach more than 14 devices to a VMUpdated hardcoded value with max data volumes limit from hypervisor capabilities.
* pr/1953:
CLOUDSTACK-9794: Unable to attach more than 14 devices to a VM
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
CLOUDSTACK-9698 [VMware] Make hardcorded wait timeout for NIC adapter hotplug as configurableJira
===
CLOUDSTACK-9698 [VMware] Make hardcoded wait timeout for NIC adapter hotplug as configurable
Description
=========
Currently ACS waits for 15 seconds (hard coded) for hot-plugged NIC in VR running on VMware to get detected by guest OS.
The time taken to detect hot plugged NIC in guest OS depends on type of VMware NIC adapter like (E1000, VMXNET3, E1000e etc.)
and guest OS itself. In uncommon scenarios the NIC detection may take longer time than 15 seconds,
in such cases NIC hotplug would be treated as failure which results in VPC tier configuration failure.
Alternatively making the wait timeout for NIC adapter hotplug as configurable will be helpful for admins in such scenarios. This is specific to VR running over VMware hypervisor.
Also in future if VMware introduces new NIC adapter types which may take time to get detected by guest OS, it is good to have flexibility of
configuring the wait timeout as fallback mechanism in such scenarios.
Fix
===
Introduce new configuration parameter (via ConfigKey) "vmware.nic.hotplug.wait.timeout" which is "Wait timeout (milli seconds) for hot plugged NIC of VM to be detected by guest OS." as fallback instead of hard coded timeout, to ensure flexibility for admins given the listed scenarios above.
Signed-off-by: Sateesh Chodapuneedi <sateesh.chodapuneedi@accelerite.com>
* pr/1861:
CLOUDSTACK-9698 Make the wait timeout for NIC adapter hotplug as configurable
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
* 4.9:
moved logrotate from cron.daily to cron.hourly for vpcrouter in cloud-early-config
CLOUDSTACK-9569: propagate global configuration router.aggregation.command.each.timeout to KVM agent
[4.9] CLOUDSTACK-9569: propagate global configuration router.aggregation.command.each.timeout to KVM agentThe router.aggregation.command.each.timeout in global configuration is only applied on new created KVM host.
For existing KVM host, changing the value will not be effective.
We need to propagate the configuration to existing host when cloudstack-agent is connected.
* pr/1856:
CLOUDSTACK-9569: propagate global configuration router.aggregation.command.each.timeout to KVM agent
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
This adds support for virtio-scsi on KVM hosts, either
for guests that are associated with a new os_type of 'Other PV Virtio-SCSI (64-bit)',
or when a VM or template is regstered with a detail parameter rootDiskController=scsi.
Update cloudstack add template dialog to allow for selecting rootDiskController with KVM
Update cloudstack kvm virtio-scsi to enable discard=unmap
Currently ACS waits for 15 seconds (hard coded) for hot-plugged NIC in VR to get detected by guest OS.
The time taken to detect hot plugged NIC in guest OS depends on type of NIC adapter like (E1000, VMXNET3, E1000e etc.)
and guest OS itself. In uncommon scenarios the NIC detection may take longer time than 15 seconds,
in such cases NIC hotplug would be treated as failure which results in VPC tier configuration failure.
Alternatively making the wait timeout for NIC adapter hotplug as configurable will be helpful for admins in such scenarios.
Also in future if VMware introduces new NIC adapter types which may take time to get detected by guest OS, it is good to have flexibility of
configuring the wait timeout as fallback mechanism in such scenarios.
Signed-off-by: Sateesh Chodapuneedi <sateesh.chodapuneedi@accelerite.com>
* 4.9:
CLOUDSTACK-9746 system-vm: logrotate config causes critical failures
CLOUDSTACK-9788: Fix exception listNetworks with pagesize=0
CLOUDSTACK-8663: Fixed various issues to allow VM snapshots and volume snapshots to exist together
Fix HVM VM restart bug in XenServer
CLOUDSTACK-9363: Fix HVM VM restart bug in XenServerHere is the longer description of the problem:
By default XenServer limits HVM guests to only 4 disks. Two of those are reserved for the ROOT disk (deviceId=0) and CD ROM (device ID=3) which means that we can only attach 2 data disks. This limit however is removed when Xentools is installed on the guest. The information that a guest has Xentools installed and can handle more than 4 disks is stored in the VM metadata on XenServer. When a VM is shut down, Cloudstack removes the VM and all the metadata associated with the VM from XenServer. Now, when you start the VM again, even if it has Xentools installed, it will default to only 4 attachable disks.
Now this problem manifests itself when you have a HVM VM and you stop and start it with more than 2 data disks attached. The VM fails to start and the only way to start the VM is to detach the extra disks and then reattach them after the VM start.
In this fix, I am removing the check which is done before creating a `VBD` which enforces this limit. This will not affect current workflow and will fix the HVM issue.
@koushik-das this is related to the "autodetect" feature that you introduced a while back (https://issues.apache.org/jira/browse/CLOUDSTACK-8826). I would love your review on this fix.
* pr/1829:
Fix HVM VM restart bug in XenServer
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
snapshots to exist together
Reverting VM to disk only snapshot in Xenserver corrupts VM
Stale NFS secondary storage on XS leads to volume creation failure from snapshot
[4.10] CLOUDSTACK-8746: VM Snapshotting implementation for KVM
* pr/977:
Fixes for testing VM Snapshots on KVM. Related to PR 977
CLOUDSTACK-8746: vm snapshot implementation for KVM
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
CLOUDSTACK-9359: IPv6 for Basic NetworkingThis PR is a proposal for adding very basic IPv6 to Basic Networking. The main goal of this PR is that the API returns a valid IPv6 address over which the Instance is reachable.
The GUI will show the IPv6 address after deployment of the Instance.
If the table VLAN has a proper IPv6 CIDR configured the DirectPodBasedNetworkGuru will calculate the IPv6 Address the Instance will obtain using EUI-64 and SLAAC: https://tools.ietf.org/search/rfc4862
In this case the _vlan_ table contained:
<pre>mysql> select * from vlan \G
*************************** 1. row ***************************
id: 1
uuid: 90e0716c-5261-4992-bb9d-0afd3006f476
vlan_id: vlan://untagged
vlan_gateway: 172.16.0.1
vlan_netmask: 255.255.255.0
description: 172.16.0.10-172.16.0.250
vlan_type: DirectAttached
data_center_id: 1
network_id: 204
physical_network_id: 200
ip6_gateway: 2001:980:7936:112::1
ip6_cidr: 2001:980:7936:112::/64
ip6_range: NULL
removed: NULL
created: 2016-07-19 20:39:41
1 row in set (0.00 sec)
mysql></pre>
It will then log:
<pre>2016-10-04 11:42:44,998 DEBUG [c.c.n.g.DirectPodBasedNetworkGuru] (Work-Job-Executor-1:ctx-1975ec54 job-186/job-187 ctx-0d967d88) (logid:275c4961) Found IPv6 CIDR 2001:980:7936:112::/64 for VLAN 1
2016-10-04 11:42:45,009 INFO [c.c.n.g.DirectPodBasedNetworkGuru] (Work-Job-Executor-1:ctx-1975ec54 job-186/job-187 ctx-0d967d88) (logid:275c4961) Calculated IPv6 address 2001:980:7936:112:4ba:80ff:fe00:e9 using EUI-64 for NIC 6a05deab-b5d9-4116-80da-c94b48333e5e</pre>
The template has to be configured accordingly:
- No IPv6 Privacy Extensions
- Use SLAAC
- Follow RFC4862
This is also described in: https://cwiki.apache.org/confluence/display/CLOUDSTACK/IPv6+in+Basic+Networking
The next steps after this will be:
- Security Grouping to prevent IPv6 Address Spoofing
- Security Grouping to filter ICMP, UDP and TCP traffic
* pr/1700:
CLOUDSTACK-676: IPv6 In -and Egress filtering for Basic Networking
CLOUDSTACK-676: IPv6 Basic Security Grouping for KVM
CLOUDSTACK-9359: IPv6 for Basic Networking with KVM
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
CLOUDSTACK-9619: Updates for SAN-assisted snapshotsThis PR is to address a few issues in #1600 (which was recently merged to master for 4.10).
In StorageSystemDataMotionStrategy.performCopyOfVdi we call getSnapshotDetails. In one such scenario, the source snapshot in question is coming from secondary storage (when we are creating a new volume on managed storage from a snapshot of ours thats on secondary storage).
This usually worked in the regression tests due to a bit of "luck": We retrieve the ID of the snapshot (which is on secondary storage) and then try to pull out its StorageVO object (which is for primary storage). If you happen to have a primary storage that matches the ID (which is the ID of a secondary storage), then getSnapshotDetails populates its Map<String, String> with inapplicable data (that is later ignored) and you dont easily see a problem. However, if you dont have a primary storage that matches that ID (which I didnt today because I had removed that primary storage), then a NullPointerException is thrown.
I have fixed that issue by skipping getSnapshotDetails if the source is coming from secondary storage.
While fixing that, I noticed a couple more problems:
1) We can invoke grantAccess on a snapshot thats actually on secondary storage (this doesnt amount to much because the VolumeServiceImpl ignores the call when its not for a primary-storage driver).
2) We can invoke revokeAccess on a snapshot thats actually on secondary storage (this doesnt amount to much because the VolumeServiceImpl ignores the call when its not for a primary-storage driver).
I have corrected those issues, as well.
I then came across one more problem:
When using a SAN snapshot and copying it to secondary storage or creating a new managed-storage volume from a snapshot of ours on secondary storage, we attach to the SR in the XenServer code, but detach from it in the StorageSystemDataMotionStrategy code (by sending a message to the XenServer code to perform an SR detach). Since we know to detach from the SR after the copy is done, we should detach from the SR in the XenServer code (without that code having to be explicitly called from outside of the XenServer logic).
I went ahead and changed that, as well.
JIRA Ticket:
https://issues.apache.org/jira/browse/CLOUDSTACK-9619
* pr/1749:
CLOUDSTACK-9619: Updates for SAN-assisted snapshots
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
This commit implements basic Security Grouping for KVM in
Basic Networking.
It does not implement full Security Grouping yet, but it does:
- Prevent IP-Address source spoofing
- Allow DHCPv6 clients, but disallow DHCPv6 servers
- Disallow Instances to send out Router Advertisements
The Security Grouping allows ICMPv6 packets as described by RFC4890
as they are essential for IPv6 connectivity.
Following RFC4890 it allows:
- Router Solicitations
- Router Advertisements (incoming only)
- Neighbor Advertisements
- Neighbor Solicitations
- Packet Too Big
- Time Exceeded
- Destination Unreachable
- Parameter Problem
- Echo Request
ICMPv6 is a essential part of IPv6, without it connectivity will break or be very
unreliable.
For now it allows any UDP and TCP packet to be send in to the Instance which
effectively opens up the firewall completely.
Future commits will implement Security Grouping further which allows controlling UDP and TCP
ports for IPv6 like can be done with IPv4.
Regardless of the egress filtering (which can't be done yet) it will always allow outbound DNS
to port 53 over UDP or TCP.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
(1) add support to create/delete/revert vm snapshots on running vms with QCOW2 format
(2) add new API to create volume snapshot from vm snapshot
(3) delete metadata of vm snapshots before stopping/migrating and recover vm snapshots after starting/migrating
(4) enable deleting of VM snapshot on stopped vm or vm snapshot is not listed in qcow2 image.
(5) enable smoke tests for vmsnaphsots on KVM
While remote executing commands/scripts through SSH in VR, ACS uses system vm keyfile.
ACS is fetching this key file using VMwareContext object which encapsulates vCenter connection handle.
This is inefficient because of dependency on getServiceContext() which means a vCenter connection
handle which is not required just to fetch a file in name space in management server.
Signed-off-by: Sateesh Chodapuneedi <sateesh.chodapuneedi@accelerite.com>
Issue
=====
While listing datacenters associated with a zone, only zone Id validation is required.
There is no need to have additional checks like zone is a legacy zone or not.
Fix
===
Removed unnecessary checks over zone ID and just checking if zone with specified ID exists or not.
Signed-off-by: Sateesh Chodapuneedi <sateesh.chodapuneedi@accelerite.com>
CLOUDSTACK-9456: Migrate master to Spring 4.xThis changes makes CloudStack use spring 4:
```
- Bump spring-framework version to 4.x and Jetty to version that runs with JDK7
- Bump servet dependency version
- Migrates various xmls to use version independent schema uris
```
Outstanding issue:
- Testing of various non-standard plugins such as network and storage plugins etc.
Since, this is a big change pinging for review -- @jburwell @karuturi @wido @murali-reddy @abhinandanprateek @DaanHoogland @GaborApatiNagy @JayapalUradi @kishankavala @K0zka @nvazquez @rafaelweingartner @pyr and others
@blueorangutan package
* pr/1638:
CLOUDSTACK-9456: Update Spring version in maven poms
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
XenServer 7 SupportThis PR adds support for XenServer 7. I have manually done the following tests
- Create a new cluster with XenServer7
- Add Primary storage: Should create an SR on XS7
- Add another XS7 host to the Pool
- Add host2 to Cloudstack
- Create VM1 from template
- Create VM2 from template
- Ping/SSH VM1 to VM2 and vice-versa
- Stop/Delete/Expunge VM2
- Create Data disk
- Attach it to VM1
- Create VM snaphsot of VM1
- Restore VM snapshot of VM1
- Delete VM snapshot of VM1
- Create Volume snapshot of Datadisk
- Create volume snapshot of Root disk
- Create new template from snapshot of root disk
- Create volume from snapshot of datadisk
- Detach datadisk volume
- Delete datadisk volume
- Aquire a public IP
- Create a static nat to VM1
- Live migrate VM1 while traffic on VM
- Delete VM1
* pr/1711:
[CLOUDSTACK-9662] Add support for XenServer 7
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Issue
====
Exception occured while creating the CPVM in the VmWare Setup using standard vswitches.
StartCommand failed due to Exception: com.vmware.vim25.AlreadyExists
message: [] com.vmware.vim25.AlreadyExistsFaultMsg: The specified key, name, or identifier already exists
Fix
===
Ensure synchronization while attempting to create port group such that simultaneous attempts are not made with same port group name on same ESXi host.
Signed-off-by: Sateesh Chodapuneedi <sateesh.chodapuneedi@accelerite.com>
- Bump spring-framework version to 4.x and Jetty to version that runs with JDK8
- Bump servet dependency version
- Migrate spring xmls to version 4, fixes schema locations that are 3.0
dependent in various xmls.
- Fix failing tests due to spring upgrade
(Thanks @marcaurele Marc-Aurèle Brothier for fixing them)
* Fix test DeploymentPlanningManagerImplTest
* Fix GloboDNS test
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
-when processing static nat rule, add a mangle table rule, to mark the traffic
from the guest vm when it has associated static nat rule so that traffic gets
routed using the route tabe of the device which has public ip associated
-fix the case where nic_device_id is empty when ip is getting disassociated
resulting in empty deviceid in ips.json
-add utility methods in CsRule, and CsRoute to add 'ip rule' and 'ip route' rules respectivley
-ensure traffic from all public interfaces are connection marked with device number, and restored
for the reverse traffic. use the connection marked number to do device specific routing table lookup
fill the device specific routing table with default route
-component tests for testing multiple public interfaces of VR
Cloudstack 9586: When using local storage with Xenserver prepareTemplate does not work with multiple primary storeThe race condition will happen whenever there are multiple primary storages and the CS tries to mount the secondary store to xenserver host simultaneously.
Due to synchronised block one mount will be successful and other thread will get the already mounted SR. Without the fix the two thread will try to mount it parallely and one will fail on Xenserver.
* pr/1765:
Cloudstack 9586: When using local storage with Xenserver prepareTemplate does not work with multiple primary store The race condition will happen whenever there are multiple primary storages and the CS tries to mount the secondary store to xenserver host simultaneously. Due to synchronised block one mount will be successful and other thread will get the already mounted SR. Without the fix the two thread will try to mount it parallely and one will fail on Xenserver.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CLOUDSTACK-9564: Fix memory leaks in VmwareContextPoolIn a recent management server crash, it was found that the largest contributor
to memory leak was in VmwareContextPool where a registry is held (arraylist)
that grows indefinitely. The list itself is not used anywhere or consumed. There
exists a hashmap (pool) that returns a list of contexts for existing poolkey
(address/username) that is used instead.
This fixes the issue by removing the arraylist registry, and limiting the
length of the context list for a given poolkey.
@blueorangutan package
* pr/1729:
CLOUDSTACK-9564: Fix memory leaks in VmwareContextPool
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
The race condition will happen whenever there are multiple primary storages and the CS tries to mount the secondary store to xenserver host simultaneously.
Due to synchronised block one mount will be successful and other thread will get the already mounted SR. Without the fix the two thread will try to mount it parallely and one will fail on Xenserver.
In a recent management server crash, it was found that the largest contributor
to memory leak was in VmwareContextPool where a registry is held (arraylist)
that grows indefinitely. The list itself is not used anywhere or consumed. There
exists a hashmap (pool) that returns a list of contexts for existing poolkey
(address/username) that is used instead.
This fixes the issue by removing the arraylist registry, and limiting the
length of the context list for a given poolkey.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CLOUDSTACK-9566 instance-id metadata for baremetal VM returns IDThere is difference in instance-id metadata across baremetal and other hypervisors.
On Baremetal
[root@ip-172-17-0-144 ~]# curl http://8.37.203.221/latest/meta-data/instance-id
6021
on Xen
[root@ip-172-17-2-103 ~]# curl http://172.17.0.252/latest/meta-data/instance-id
cbeb517a-e833-4a0c-b1e8-9ed70200fbbf
In both cases it should be vm's uuid.
* pr/1738:
CLOUDSTACK-9566 instance-id metadata for baremetal VM returns ID
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
This commit adds a additional VirtIO channel with the name
'org.qemu.guest_agent.0' to all Instances.
With the Qemu Guest Agent the Hypervisor gains more control over the Instance if
these tools are present inside the Instance, for example:
* Power control
* Flushing filesystems
* Fetching Network information
In the future this should allow safer snapshots on KVM since we can instruct the
Instance to flush the filesystems prior to snapshotting the disk.
More information: http://wiki.qemu.org/Features/QAPI/GuestAgent
Keep in mind that on Ubuntu AppArmor still needs to be disabled since the default
AppArmor profile doesn't allow libvirt to write into /var/lib/libvirt/qemu
This commit does not add any communication methods through API-calls, it merely
adds the channel to the Instances and installs the Guest Agent in the SSVMs.
With the addition of the Qemu Guest Agent channel a second channel appears in /dev
on a SSVM as a VirtIO port.
The order in which the ports are defined in the XML matters for the naming inside
the SSVM VM and by not relying on /dev/vportXX but looking for a static name the
SSVM still boots properly if the order in the XML definition is changed.
A SSVM with both ports attached will have something like this:
root@v-215-VM:~# ls -l /dev/virtio-ports
total 0
lrwxrwxrwx 1 root root 11 May 13 21:41 org.qemu.guest_agent.0 -> ../vport0p2
lrwxrwxrwx 1 root root 11 May 13 21:41 v-215-VM.vport -> ../vport0p1
root@v-215-VM:~# ls -l /dev/vport*
crw------- 1 root root 251, 1 May 13 21:41 /dev/vport0p1
crw------- 1 root root 251, 2 May 13 21:41 /dev/vport0p2
root@v-215-VM:~#
In this case the SSVM port points to /dev/vport0p1, but if the order in the XML
is different it might point to /dev/vport0p2
By looking for a portname with a pre-defined pattern in /dev/virtio-ports we
do not rely on the order in the XML definition.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
CLOUDSTACK-9379: Support nested virtualization at VM level on VMware Hypervisor## Introduction
[JIRA TICKET](https://issues.apache.org/jira/browse/CLOUDSTACK-9379)
It is desired to support nested virtualization at VM level for VMware hypervisor. Current behaviour supports enabling/desabling global nested virtualization by modifying global config `'vmware.nested.virtualization'`. It is wished to improve this feature, having control at VM level instead of a global control only.
A new global configuration is added, to enable/disable VM nested virtualization control: `'vmware.nested.virtualization.perVM'`. Default value=false
After a vm deployment or start command, vm params include `'nestedVirtualizationFlag'` key and its value is:
- true -> nested virtualization enabled
- false -> nested virtualization disabled
**We will determinate nested virtualization enabled/disabled by examining this 3 values:**
- **(1)** global configuration `'vmware.nested.virtualization'` value
- **(2)** global configuration `'vmware.nested.virtualization.perVM'` value
- **(3)** `'nestedVirtualizationFlag'` value in `user_vm_details` if present, `null` if not.
Using this 3 values, there are different use cases:
- **(1)** = TRUE, **(2)** = TRUE, **(3)** is null -> _ENABLED_
- **(1)** = TRUE, **(2)** = TRUE, **(3)** = TRUE -> _ENABLED_
- **(1)** = TRUE, **(2)** = TRUE, **(3)** = FALSE -> _DISABLED_
- **(1)** = TRUE, **(2)** = FALSE, **(3)** indifferent -> _ENABLED_
- **(1)** = FALSE, **(2)** = TRUE, **(3)** is null -> _DISABLED_
- **(1)** = FALSE, **(2)** = TRUE, **(3)** = TRUE -> _ENABLED_
- **(1)** = FALSE, **(2)** = TRUE, **(3)** = FALSE -> _DISABLED_
- **(1)** = FALSE, **(2)** = FALSE, **(3)** indifferent -> _DISABLED_
* pr/1542:
CLOUDSTACK-9379: Support nested virtualization at VM level on VMware Hypervisor
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CLOUDSTACK-9491: incorrect parsing of device list to find ethernet index of plugged NICIn VmwareResource, findRouterEthDeviceIndex() method find ethernet interface index given
the mac address. This method is used, once a nic is plugged to determine ethernet interface.
"/proc/sys/net/ipv4/conf" from the VR and looped through the devices to find the right
ethernet interface. Howver current logic read it once, and loops through the device list.
Its observerd device may not show up '/proc/sys/net/ipv4/conf' immediatly once NIC is plugged
in the VM from vCenter.
Fix ensured, while waiting for 15 sec in the loop, read the latest content from /proc/sys/net/ipv4/conf
, so that right device list is processed.
Manual tested VPC scenarios of adding new tiers which uses findRouterEthDeviceIndex, to find the guest/public network ethernet index.
* pr/1681:
CLOUDSTACK-9491: incorrect parsing of device list to find ethernet index of plugged NIC
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CLOUDSTACK-9503: Increased the VR script timeout. Most of the changes are about converting int/long time values to joda Duration.
* pr/1745:
CLOUDSTACK-9503: Increased the VR script timeout. Most of the changes are about converting int/long time values to joda Duration.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CLOUDSTACK-9502: DS template copies dont get deleted in VMware ESXi with multiple clusters and zone wide storage (include CLOUDSTACK-9386 into 4.9 release branch)Include #1560 into 4.9 release branch
* pr/1676:
CLOUDSTACK-9502: DS template copies don’t get deleted in VMware ESXi with multiple clusters and zone wide storage
Signed-off-by: John Burwell <meaux@cockamamy.net>
* 4.9:
CLOUDSTACK-8830: Fix for vm snapshots in Vmware, could not create vm snapshot until 12 minutes after vm creation due to vCenter sent null name on snpashot recent task
Made the changes to improve logging.CLOUSTACK-9465 Several log refactoring/improvement suggestions.
There are two scenarios of logging which needs refactoring/improvement:
Method invocation replaced by variable
This means that in the logging code, the method invocation is pre-defined as a variable. for simplicity, the method invocation should be replaced by the variable.
Delete variable which must be null
The variable in the logging code is null, there is no need to put the variable there.
* pr/1705:
Made the changes to improve logging.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
This PR adds an ability to Pass a new parameter, locationType,
to the “createSnapshot” API command. Depending on the locationType,
we decide where the snapshot should go in case of managed storage.
There are two possible values for the locationType param
1) `Standard`: The standard operation for managed storage is to
keep the snapshot on the device. For non-managed storage, this will
be to upload it to secondary storage. This option will be the
default.
2) `Archive`: Applicable only to managed storage. This will
keep the snapshot on the secondary storage. For non-managed
storage, this will result in an error.
The reason for implementing this feature is to avoid a single
point of failure for primary storage. Right now in case of managed
storage, if the primary storage goes down, there is no easy way
to recover data as all snapshots are also stored on the primary.
This features allows us to mitigate that risk.
In VmwareResource, findRouterEthDeviceIndex() method find ethernet interface index given
the mac address. This method is used, once a nic is plugged to determine ethernet interface.
"/proc/sys/net/ipv4/conf" from the VR and looped through the devices to find the right
ethernet interface. However current logic read it once, and loops through the device list.
Its observerd device may not show up '/proc/sys/net/ipv4/conf' immediatly once NIC is plugged
in the VM from vCenter.Fix ensured, while waiting for 15 sec in the loop, read the latest
content from /proc/sys/net/ipv4/conf, so that right device list is processed.
By adding a Random Number Generator device to Instances we can prevent
entropy starvation inside guest.
The default source is /dev/random on the host, but this can be configured
to another source when present, for example a hardware RNG.
When enabled it will add the following to the Instance's XML definition:
<rng model='virtio'>
<rate period='1000' bytes='2048' />
<backend model='random'>/dev/random</backend>
</rng>
If the Instance has the proper support, which most modern distributions have,
it will have a /dev/hwrng device which it can use for gathering entropy.
More information: https://libvirt.org/formatdomain.html#elementsRng
CLOUDSTACK-9422: Granular 'vmware.create.full.clone' as Primary Storage setting### Introduction
For VMware, It is possible to decide creating VMs as full clones on ESX HV, adjusting `vmware.create.full.clone` global setting. We would like to introduce this property as a primary storage detail, and use its value instead of global setting's value.
We propose introducing `fullCloneFlag` on `PrimaryDataStoreTO` sent on `CopyCommand`. This way we can reconfigure `VmwareStorageProcessor` and `VmwareStorageSubsystemCommandHandler` similar as it was done for `nfsVersion` but refactoring it to be more general.
* pr/1602:
CLOUDSTACK-9422: Granular VMware vms creation as full clones on HV
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
On snapshot backup, this converts the rbd raw format on disk to qcow2 for compression.
* pr/1645:
CLOUDSTACK-9461
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
Adding support for cross-cluster storage migration for managed storage when using XenServerThis PR adds support for cross-cluster storage migration of VMs that make use of managed storage with XenServer.
Managed storage is when you have a 1:1 mapping between a virtual disk and a volume on a SAN (in the case of XenServer, an SR is placed on this SAN volume and a single virtual disk placed in the SR).
Managed storage allows features such as storage QoS and SAN-side snapshots to work (sort of analogous to VMware VVols).
This PR focuses on enabling VMs that are using managed storage to be migrated across XenServer clusters.
I have successfully run the following tests on this branch:
TestVolumes.py
TestSnapshots.py
TestVMSnapshots.py
TestAddRemoveHosts.py
TestVMMigrationWithStorage.py (which is a new test that is being added with this PR)
* pr/1671:
Adding support for cross-cluster storage migration for managed storage when using XenServer
Signed-off-by: Rajani Karuturi <rajani.karuturi@accelerite.com>
CLOUDSTACK-9428: Fix for CLOUDSTACK-9211 - Improve performance of 3D GPU support in cloud-plugin-hypervisor-vmwareJIRA TICKET: https://issues.apache.org/jira/browse/CLOUDSTACK-9428
### Introduction
On #1310 passing vRAM size to support 3D GPU problem was addressed on VMware. It was found out that it could be improved to increase performance by reducing extra API calls, as we'll describe later
### Improvement
On WMware, `VmwareResource` manages execution of `StartCommand.` Before sending power on command to ESXi hypervisor, vm is configured by calling `reconfigVMTask` web method on vSphere's client `VimPortType` web service.
It was found out that we were using this method 2 times when passing vRAM size, as it implied creating a new vm config spec only editing video card specs and making an extra call to `reconfigVMTask.`
We propose reducing the extra web service call by adjusting vm's config spec. This way video card gets properly configured (when passing vRAM size) in the same configure call, increasing performance.
### Use case (passing vRAM size)
* Deploy a new VM, let its id be X
* Stop VM
* Execute SQL, where X is vm's id and Z is vRAM size (in kB):
````
INSERT INTO cloud.user_vm_details (vm_id, name, value) VALUES (X, 'mks.enable3d', 'true');
INSERT INTO cloud.user_vm_details (vm_id, name, value) VALUES (X, 'mks.use3dRenderer', 'automatic');
INSERT INTO cloud.user_vm_details (vm_id, name, value) VALUES (X, 'svga.autodetect', 'false');
INSERT INTO cloud.user_vm_details (vm_id, name, value) VALUES (X, 'svga.vramSize', Z);
````
* Start VM
* pr/1605:
CLOUDSTACK-9428: Add marvin test
CLOUDSTACK-9428: Fix for CLOUDSTACK-9211 - Improve performance
Signed-off-by: Rafael Weingärtner <rafael@apache.org>
Often, patch and security releases do not require schema migrations or
data migrations. However, if an empty upgrade class and associated
scripts are not defined, the upgrade process will break. With this
change, if a release does not have an upgrade, a noop DbUpgrade is added
to the upgrade path. This approach allows the upgrade to proceed and
for the database to properly reflect the installed version. This change
should make the release process simpler as RMs no longer need to
rememeber to create this boilerplate code when starting a new release.
Beginning with the 4.8.2.0 and 4.9.1.0 releases, the project will
formally adopt a four (4) position release number to properly accomodate
rekeases that contain only CVE fixes. The DatabaseUpgradeChecker and
Version classes made assumptions that they would always parse and
compare three (3) position version numbers. This change adds the
CloudStackVersion value object that supports both three (3) and four (4)
version numbers. It encapsulates version comparsion logic, as well as,
the rules to allow three (3) and four (4) to interoperate.
* Modifies DatabaseUpgradeChecker to handle derive an upgrade path for
a version that was not explicitly specified. It determines the
releases the first release before it with database migrations and uses
that list as the basis for the list for version being calculated. A
noop upgrade is then added to the list which causes no schema changes
or data migrations, but will update the database to the version.
* Adds unit tests for the upgrade path calculation logic in
DatabaseUpgradeChecker
* Removes dummy upgrade logic for the 4.8.2.0 introduced in previous
versions of this patch
* Introduces the CloudStackVersion value object which parses and
compares three (3) and four (4) position version numbers. This class
is intended to replace com.cloud.maint.Version.
* Adds the junit-dataprovider dependency -- allowing test data to be
concisely generated separately from the execution of a test case.
Used extensively in the CloudStackVersionTest.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
[blocker] CLOUDSTACK-9452: add python-argparse dependency on el6,7 rpmsThe patchviasocket script was rewritten in Python from PR #1533 and made
assumptions that Python 2.7 would be available. In case of CentOS, python 2.7
may not be available or installed. This change ensures that python-argparse
is installed which is used by this script.
/cc @wido @sverrirab @karuturi @jburwell
@blueorangutan package
* pr/1634:
CLOUDSTACK-9452: add python-argparse dependency on el6,7 rpms
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
The patchviasocket script was rewritten in Python from PR #1533 and made
assumptions that Python 2.7 would be available. In case of CentOS, python 2.7
may not be available or installed. This change ensures that python-argparse
is installed which is used by this script.
Expose cmd error in the logs when patch command fails.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
We noticed that when an exception occurs within the cleanup loop inside
the deletePhysicalDisk routine that the previously allocated contexts
are not cleaned up. This seemed to cause an eventual crash of the host
agent after multiple exceptions within the loop.
In addition to ensuring the contexts are always freed we also improved
the logging when exceptions do occur to include the actual return code
from the underlying library in deletePhysicalDisk and deleteSnapshot.
* 4.8:
CLOUDSTACK-9353: [XenServer] Fixed VM migration with storage
Added ASF license to unit test file
Added unit test to verify ordering
Fixed ordering of network ACL rules being sent to the VR. The comparator was inverted
* 4.7:
CLOUDSTACK-9353: [XenServer] Fixed VM migration with storage
Added ASF license to unit test file
Added unit test to verify ordering
Fixed ordering of network ACL rules being sent to the VR. The comparator was inverted
CLOUDSTACK-9368: Fix for Support configurable NFS version for Secondary Storage mounts## Description
JIRA TICKET: https://issues.apache.org/jira/browse/CLOUDSTACK-9368
This pull request address a problem introduced in #1361 in which NFS version couldn't be changed after hosts resources were configured on startup (for hosts using `VmwareResource`), and as host parameters didn't include `nfs.version` key, it was set `null`.
## Proposed solution
In this proposed solution `nfsVersion` would be passed in `NfsTO` through `CopyCommand` to `VmwareResource`, who will check if NFS version is still configured or not. If not, it will use the one sent in the command and will set it to its storage processor and storage handler. After those setups, it will proceed executing command.
* pr/1518:
CLOUDSTACK-9368: Fix for Support configurable NFS version for Secondary Storage mounts
Signed-off-by: Will Stevens <williamstevens@gmail.com>
Convert patchviasocket to python (removes perl dependency for KVM agent)As requested here: https://github.com/apache/cloudstack/pull/1495
No scripts are using perl so that install requirement can be removed.
The new scripts are using standard python packages only.
Includes extensive unit test.
Note: perl-modules requirement is missing (fixed in mentioned PR) so do not merge that onto master.
* pr/1533:
Revert "Add perl-modules as install dependency for cloudstack-agent"
patchviasocket improve error handling
Convert patchviasocket to python (removes perl dependency for KVM agent)
Signed-off-by: Will Stevens <williamstevens@gmail.com>
As requested here: https://github.com/apache/cloudstack/pull/1495
No scripts are using perl so that install requirement can be removed.
The new scripts are using standard python packages only.
Includes extensive unit test.
Taking fast and efficient volume snapshots with XenServer (and your storage provider)A XenServer storage repository (SR) and virtual disk image (VDI) each have UUIDs that are immutable.
This poses a problem for SAN snapshots, if you intend on mounting the underlying snapshot SR alongside the source SR (duplicate UUIDs).
VMware has a solution for this called re-signaturing (so, in other words, the snapshot UUIDs can be changed).
This PR only deals with the CloudStack side of things, but it works in concert with a new XenServer storage manager created by CloudOps (this storage manager enables re-signaturing of XenServer SR and VDI UUIDs).
I have written Marvin integration tests to go along with this, but cannot yet check those into the CloudStack repo as they rely on SolidFire hardware.
If anyone would like to see these integration tests, please let me know.
JIRA ticket: https://issues.apache.org/jira/browse/CLOUDSTACK-9281
Here's a video I made that shows this feature in action:
https://www.youtube.com/watch?v=YQ3pBeL-WaA&list=PLqOXKM0Bt13DFnQnwUx8ZtJzoyDV0Uuye&index=13
* pr/1403:
Faster logic to see if a cluster supports resigning
Support for backend snapshots with XenServer
Signed-off-by: Will Stevens <williamstevens@gmail.com>
CPU socket count reporting correctionCPU socket count reporting correction
From https://github.com/MissionCriticalCloud/cosmic-plugin-hypervisor-kvm/pull/16
* pr/1520:
Remove empty spaces causing the build to fail
CPU socket count reporting correction
Signed-off-by: Will Stevens <williamstevens@gmail.com>
CLOUDSTACK-8800 : Improved the listVirtualMachines API call to include memory utilization information for a VMThis PR introduces the changes proposed in PR #780 with some work to make the code null safe.
During this PR, I have also removed some unused code.
* pr/1444:
Removed unnecessary check when creating the “userVmResponse” object.
Fixed issues from CLOUDSTACK-8800 that were introduced in PR 780
CLOUDSTACK-8800 : Improved the listVirtualMachines API call to include memory utilization information for a VM for xenserver,kvm and for vmware.
Signed-off-by: Will Stevens <williamstevens@gmail.com>
This feature allows root administrators to define new roles and associate API
permissions to them.
A limited form of role-based access control for the CloudStack management server
API is provided through a properties file, commands.properties, embedded in the
WAR distribution. Therefore, customizing API permissions requires unpacking the
distribution and modifying this file consistently on all servers. The old system
also does not permit the specification of additional roles.
FS:
https://cwiki.apache.org/confluence/display/CLOUDSTACK/Dynamic+Role+Based+API+Access+Checker+for+CloudStack
DB-Backed Dynamic Role Based API Access Checker for CloudStack brings following
changes, features and use-cases:
- Moves the API access definitions from commands.properties to the mgmt server DB
- Allows defining custom roles (such as a read-only ROOT admin) beyond the
current set of four (4) roles
- All roles will resolve to one of the four known roles types (Admin, Resource
Admin, Domain Admin and User) which maintains this association by requiring
all new defined roles to specify a role type.
- Allows changes to roles and API permissions per role at runtime including additions or
removal of roles and/or modifications of permissions, without the need
of restarting management server(s)
Upgrade/installation notes:
- The feature will be enabled by default for new installations, existing
deployments will continue to use the older static role based api access checker
with an option to enable this feature
- During fresh installation or upgrade, the upgrade paths will add four default
roles based on the four default role types
- For ease of migration, at the time of upgrade commands.properties will be used
to add existing set of permissions to the default roles. cloud.account
will have a new role_id column which will be populated based on default roles
as well
Dynamic-roles migration tool: scripts/util/migrate-dynamicroles.py
- Allows admins to migrate to the dynamic role based checker at a future date
- Performs a harder one-way migrate and update
- Migrates rules from existing commands.properties file into db and deprecates it
- Enables an internal hidden switch to enable dynamic role based checker feature
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CLOUDSTACK-9350: KVM-HA- Fix CheckOnHost for Local storage- KVM-HA- Fix CheckOnHost for Local storage
- Also skip HA on VMs that are using local storage
* pr/1496:
CLOUDSTACK-9350: KVM-HA- Fix CheckOnHost for Local storage - Also skip HA on VMs that are using local storage
Signed-off-by: Will Stevens <williamstevens@gmail.com>
CID-1338387: Deletion of method endPointSelector.selectHypervisorHostFollowing the discussions and analysis presented on PR #1056 create by @DaanHoogland
This PR is intended to push those changes that were discussed there regarding the of endPointSelector.selectHypervisorHost method.
* pr/1124:
Deletion of method endPointSelector.selectHypervisorHost
Signed-off-by: Will Stevens <williamstevens@gmail.com>
CLOUDSTACK-8302: Removing snapshots on RBDSnapshot removing implemented if primary datastore is RBD
https://issues.apache.org/jira/browse/CLOUDSTACK-8302
* pr/1230:
CLOUDSTACK-8302 - Cleanup snapshot on KVM with RBD Snapshot removing implemented on RBD. 1. On management side: when created new shanpshot we checking if our primary storage is RBD, then do not remove record from cloud.snapshot_store_ref with link to Ceph image via 'install_path' field. 2. On management side: when removing snapshot, also send command to agent 'DeleteCommand'. 3. On agent side: method implemented 'public Answer deleteSnapshot(final DeleteCommand cmd)'
Signed-off-by: Will Stevens <williamstevens@gmail.com>
It was worked around some possible runtime exceptions introduced by the
changes that were added by the PR 780. Basically, the points in which a
null pointer exception could happen, we added safety checks to avoid
them. It was create a specific method do that, all together test cases
were created for this newly method that was added.
[4.7] vmware: improve support for disks- Improve disk chain usage while attaching, migrating disks
- Gets root disk controller based diskDeviceBusName from volume's chain info
* pr/1365:
vmware: improve support for disks
Signed-off-by: Will Stevens <williamstevens@gmail.com>
4.9 mvn version safeupgradeonlyUpgrades maven dependencies versions that can be safely upgraded without breaking console-proxy/crypto usage.
Bisected changes from: https://github.com/apache/cloudstack/pull/1397
cc @swill @DaanHoogland
* pr/1510:
maven: fix dependency version support by JDK7
further maven dependency updates from Daan
framework/quota: fix checkstyle issue
maven: Upgrade dependency versions
Signed-off-by: Will Stevens <williamstevens@gmail.com>
CLOUDSTACK-9142 Migrate VM changes xmlDesc in a safe wayThe problem arises when the origin hypervisor has an ip addres that ends with 1, like '10.10.10.1' and the qemu VM description is containing an address that has that as part of its address, '10.10.10.100' for instance.
now migrating to '10.10.10.10' will change both addresses in the xml description file for qemu. It is fixed and unit tests are added. I am not sure yet how to integration test this. Regression will probably work so creating a PR now.
* pr/1348:
CLOUDSTACK-9142 Migrate VM changes xmlDesc in a safe way
Signed-off-by: Will Stevens <williamstevens@gmail.com>
Snapshot removing implemented on RBD.
1. On management side: when created new shanpshot we checking if our primary storage is RBD,
then do not remove record from cloud.snapshot_store_ref with link to Ceph
image via 'install_path' field.
2. On management side: when removing snapshot, also send command to agent 'DeleteCommand'.
3. On agent side: method implemented 'public Answer deleteSnapshot(final DeleteCommand cmd)'
CLOUDSTACK-9130: Make RebootCommand similar to start/stop/migrate agent commands w.r.t. "execute in sequence" flag
RebootCommand now behaves in the same way as start/stop/migrate agent commands w.r.t. to sequential/parallel execution.
* pr/1200:
CLOUDSTACK-9130: Make RebootCommand similar to start/stop/migrate agent commands w.r.t. "execute in sequence" flag RebootCommand now behaves in the same way as start/stop/migrate agent commands w.r.t. to sequential/parallel execution.
Signed-off-by: Will Stevens <williamstevens@gmail.com>
Removed unnecessary code from getGuestOsType in CitrixResourceBaseConsidering that all mapping of Guest OS Names to their respective hypervisor compatible types is made thorugh accessing a database, we've decided to remove a bit of code in the XcpOssResource class which was doing that same thing for 2 different OS's (both of which ARE in the database). That has led us to a bunch of unused parameters in the getGuestOsType method from its superclass, which we've also decided to remove. Test cases were added for four different possibilities for the platformEmulator String: one for a null String, one for a blank String, one for an empy String and one for a random case with a valid String.
* pr/1262:
Remove test cases duplicated code.
Removed unnecessary code from getGuestOsType in CitrixResourceBase
Signed-off-by: Will Stevens <williamstevens@gmail.com>
- Improve disk chain usage while attaching, migrating disks
- Gets root disk controller based diskDeviceBusName from volume's chain info
- Refactor and move VirtualMachineDiskInfo to cloud-utils
- Allows mixing of scsi controller types
- Fixes a NPE case with map passed as null, for example in case of detach volume
command
- Use a osdefault translator that allow use of recent os types added (enums of
which) are not available in the sdk
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
SecurityGroupRulesCmd code cleanupWrote a test and cleaned some duplicate code with the objective to evaluate the jenkins pull request process at builds.a.o
worthwhile to keep, IMHO.
* pr/1287:
SecurityGroupRulesCmd code cleanup review comments handled
deal with PMD warnings
code cleanup
security rules test
remove autogenerated pydev files
Signed-off-by: Koushik Das <koushik@apache.org>
CLOUDSTACK-8800 : Improved the listVirtualMachines API call to include memory utilization information for a VMfor xenserver,kvm and for vmware.
* pr/780:
CLOUDSTACK-8800 : Improved the listVirtualMachines API call to include memory utilization information for a VM for xenserver,kvm and for vmware.
Signed-off-by: Kishan Kavala <kishan@apache.org>
CLOUDSTACK-9252: Support configurable NFS version for Secondary Storage mountsJIRA Ticket: https://issues.apache.org/jira/browse/CLOUDSTACK-9252
### Description of the problem
After starting secondary storage VM, secondary storage tries to be mounted but fails with error: <code>Protocol family not supported</code>
It was found out that adding <code>-o vers=X</code> to mount command it would work, where <code>X</code> is the desired NFS version to use.
If it is desired to mount a store with a specific NFS version, it has passed in <code>image_store_details</code> table for a store with id <code>Y</code> as a property:
| store_id| name| value |
|:-------------:|:-------------:|:-------------:|
|Y|nfs.version|X (e.g. 3)|
Where X stands for NFS version
* pr/1361:
CLOUDSTACK-9252: Last refactor, passing nfs version to ssvm
CLOUDSTACK-9252: Add missing licence header
CLOUDSTACK-9252: Little refactor
CLOUDSTACK-9252: Mock application context for unit test
CLOUDSTACK-9252: Add unit tests
CLOUDSTACK-9252: New refactor
CLOUDSTACK-9252: Remove static dependencies, refactor
CLOUDSTACK-9252: Remove duplicates getNfsVersion, refactor
CLOUDSTACK-9252: Support configurable nfs version
CLOUDSTACK-9252: Add nfs version to commands
Signed-off-by: Rafael Weingärtner <rafael@apache.org>
CLOUDSTACK-9211: Support passing vRAM size to support 3D GPU on VmwareJIRA TICKET:
https://issues.apache.org/jira/browse/CLOUDSTACK-9211
CS support passing hypervisor options by creating entries in <code>vm_template_details</code> or <code>user_vm_details</code>
To enable software 3D GPU 4 options needs to be added:
| name| value |
|:-------------:|:-------------:|
|mks.enable3d|true|
|mks.use3dRenderer|automatic|
|svga.autodetect|false|
|svga.vramSize|(size in KB) e.g. 131072|
Currently all options except <code>svga.vramSize</code> works, VMs always get configured with default 64Mb videoRAM instead of the one provided on the option.
* pr/1310:
CLOUDSTACK-9211: Unit test for 3dgpu support
CLOUDSTACK-9211: Refactor vm vram size setter method
CLOUDSTACK-9211: Add javadoc and refactor method
CLOUDSTACK-9211: Support passing vRAM size to support 3D GPU on Vmware
Signed-off-by: Rafael Weingärtner <rafael@apache.org>
Fixed return type Void to void in DataMotionStrategy.The main changes are:
- Changing methods Void to void.
- Removal of the method Void copyAsync(DataObject srcData, DataObject
destData, AsyncCompletionCallback<CopyCommandResult> callback) that was
never used.
We noticed that methods form that class are using the return type Void
with capital V. This way that method has to return a null value at the
end.
Removed trim lines from XenServerStorageMotionStrategy.
The trim lines were removed from XenServerStorageMotionStrategy.
* pr/969:
Changed return of methods from DataMotionStrategy, Void to void
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
test: Fix Libvirt test so that it works on WindowsThis test failed on Windows, using the File.separator it should run fine on Windows.
* pr/1242:
test: Fix Libvirt test so that it works on Windows
Signed-off-by: Remi Bergsma <github@remi.nl>
* 4.7:
Fix execution counter to support separate counts per thread
Add test to check that each thread has it's own execution counter
CLOUDSTACK-9231: Root volume migration from one primary to another primary storage within the same cluster is failing
This situation arises when there are two management server accessing the same database.
When the migration request comes the command is forwarded from one management server to another because
the host is owned by the second management server. So, serialization of map from one to another fails.
This is fixed by converting the maps to lists.
CLOUDSTACK-9127 Missing PV-bootloader-args for "SUSE Linux Enterprise Server 10 SP2 and SP3"
ISSUE
--------
STOP-START of SUSE Linux VMs fail, as PV-bootloader-args are missing during the start command.
DESCRIPTION
----------------------
Repro steps
1. Upload Suse ISO
2. Create a VM with this ISO, and install it.
3. Detach ISO from the VM.
4. Reboot the VM, :>>>> This will work fine, as the pv-bootloader-args are not missing during reboot.
5.Stop the VM from CCP(VM will get destroyed in Xencenter)
6. Start the same VM from CCP , it will try to start but will fail.
Before Applying the fix
--------------------------------
Before applying the starting the VM failed with following exception
com.cloud.utils.exception.CloudRuntimeException: Unable to start VM(i-2-6-VM) on host(7cfd6388-b763-4c09-b3a3-9679db2904a3) due to Task failed! Task record: uuid: 21a6799f-9523-7c0e-bb86-1de750a38d74
nameLabel: Async.VM.start_on
nameDescription:
allowedOperations: []
currentOperations: {}
created: Wed Dec 09 07:00:29 UTC 2015
finished: Wed Dec 09 07:00:31 UTC 2015
status: failure
residentOn: com.xensource.xenapi.Host@513d238c
progress: 1.0
type: <none/>
result:
errorInfo: [BOOTLOADER_FAILED, OpaqueRef:0b10b6ac-837d-29af-da9d-6ef1e11a064a, Unable to find partition containing kernel
]
otherConfig: {}
subtaskOf: com.xensource.xenapi.Task@aaf13f6f
subtasks: []
After Applying the fix
--------------------------
After applying the fix I am able to start the vm.
* pr/1196:
CLOUDSTACK-9127 Missing PV-bootloader-args for "SUSE Linux Enterprise Server 10 SP2 and SP3"
Signed-off-by: Daan Hoogland <daan@onecht.net>
* 4.6:
[UI] bug fix: Delete added ACL lists is not available for Domain Admin and normal users
CLOUDSTACK-4787: Allow users to select disk controller for VM/template
CLOUDSTACK-4787 Allow selection of scsi controller type in vSphere
[4.6] CLOUDSTACK-4787 - vmware diskcontrollersSame as #1131 (see this for screenshots etc)
* pr/1132:
CLOUDSTACK-4787: Allow users to select disk controller for VM/template
CLOUDSTACK-4787 Allow selection of scsi controller type in vSphere
Signed-off-by: Daan Hoogland <daan@onecht.net>
CLOUDSTACK-9095 : Hypervisor changes to support UserData for Nuage VSP
* pr/1142:
CLOUDSTACK-9095 : Hypervisor changes to support UserData for Nuage VSP
Signed-off-by: Remi Bergsma <github@remi.nl>
Quota service while allowing for scalability will make sure that the cloud is
not exploited by attacks, careless use and program errors. To address this
problem, we propose to employ a quota-enforcement service that allows resource
usage within certain bounds as defined by policies and available quotas for
various entities. Quota service extends the functionality of usage server to
provide a measurement for the resources used by the accounts and domains using a
common unit referred to as cloud currency in this document. It can be configured
to ensure that your usage won’t exceed the budget allocated to accounts/domain
in cloud currency. It will let user know how much of the cloud resources he is
using. It will help the cloud admins, if they want, to ensure that a user does
not go beyond his allocated quota. Per usage cycle if a account is found to be
exceeding its quota then it is locked. Locking an account means that it will not
be able to initiat e a new resource allocation request, whether it is more
storage or an additional ip. Needless to say quota service as well as any action
on the account is configurable.
Changes from Github code review:
- Added marvin test for quota plugin API
- removed unused commented code
- debug messages in debug enabled check
- checks for nulls, fixed access to member variables and feature
- changes based on PR comments
- unit tests for UsageTypes
- unit tests for all Cmd classes
- unit tests for all service and manager impls
- try-catch-finally or try-with-resource in dao impls for failsafe db switching
- remove dead code
- add missing quota calculation case (regression fixed)
- replace tabs with spaces in pom.xmls
- quota: though default value for quota_calculated is 0, the usage server
makes it null while entering usage entries. Flipping the condition so
as to acocunt for that.
- quotatypes: fix NPE in quota type
- quota framework test fixes
- made statement period configurable
- changed default email templates to reflect the fact that exhausted quota may not result in a locked account
- added quotaUpdateCmd that refreshes quota balances and sends alerts and statements
- report quotaSummary command returns quota balance, quota usage and state for all account
- made UI framework changes to allow for text area input in edit views
- process usage entries that have greater than 0 usage
- orocess quota entries only if tariff is non zero
- if there are credit entries but no balance entry create a dummy balance entry
- remove any credit entries that are before the last balance entry
when displaying balance statement
- on a rerun the last balance is now getting added
FS: https://cwiki.apache.org/confluence/display/CLOUDSTACK/Quota+Service+-+FS
PR: https://github.com/apache/cloudstack/pull/768
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CLOUDSTACK-9025: Fixed can't create usable template from snapshot in Xenserver and Vmwarehttps://issues.apache.org/jira/browse/CLOUDSTACK-9025
Fix also reverts below commit as below solution making assumption about hypervisor which are not applicable in case of XenServer and VmWare
Revert "CLOUDSTACK-8964: Can't create template or volume from snapshot"
This reverts commit ccf5d75cfb.
Testing:
Able to deploy VM successfully from template created from linked clone snapshot on XenServer.
* pr/1176:
CLOUDSTACK-9025: Fixed can't create usable template from snapshot in Xenserver and Vmware
Signed-off-by: Daan Hoogland <daan@onecht.net>
CLOUDSTACK-9047 rename enumsmake enums adhere to best practice naming conventions
* pr/1049:
CLOUDSTACK-9046 rename enums to adhere to naming conventions
CLOUDSTACK-9046 renamed enums in kvm plugin
CLOUDSTACK-9047 use 'State's only with context there are more types called 'State' (or to be called so but now 'state') So remove imports and prepend their enclosing class/context to them.
Signed-off-by: Daan Hoogland <daan@onecht.net>
Removed the PlannerBase class because it is does not bring contribution to ACS' code.Removed the PlannerBase class because it is does not bring contribution to ACS' code.
We changed com.cloud.deploy.FirstFitPlanner, now it doesnt extends the PlannerBase and implements the com.cloud.deploy.DeploymentPlanner.
We also removed the method com.cloud.deploy.DeploymentPlanner.check(VirtualMachineProfile, DeploymentPlan, DeployDestination, ExcludeList) that was not used anywhere.
Additionally, we removed the _ from FirstFitPlanner's attributes name, in order to have them in a more standard way.
* pr/1108:
Removed PlannerBase empty class
Signed-off-by: Remi Bergsma <github@remi.nl>
Fix also reverts below commit as below solution making assumption about hypervisor which are not applicable
in case of XenServer and VmWare
Revert "CLOUDSTACK-8964: Can't create template or volume from snapshot"
This reverts commit ccf5d75cfb.
Add function to set vapic, spinlock and retries
Add function to get retry value
Modify toString to output appropriate XML for spinlock value if set
CLOUDSTACK-9004: Add features to HyperVEnlightenmentFeatureDef
Refactored set methods to get rid of code duplication.
Modified unit tests accordingly
Create test cases to getPatchFilePath method and class names changed In this commit we created tests cases for the respective classes in package com.cloud.hypervisor.xenserver.resource.
We added test cases to check the implementation of com.cloud.hypervisor.xenserver.resource.CitrixResourceBase.getPatchFiles. Therefore, we test in a more comprehensive way the tests that already exist to check the code of com.cloud.hypervisor.xenserver.resource.CitrixResourceBase.getPatchFilePath.
We added a new abstract class, called com.cloud.hypervisor.xenserver.resource.CitrixResourceBaseTest.java
This class has two tests methods:
* com.cloud.hypervisor.xenserver.resource.CitrixResourceBaseTest.testGetPathFilesExeption(CitrixResourceBase), this method tests if the getPatchFilePath() method throws the com.cloud.utils.exception.CloudRuntimeException.CloudRuntimeException when the com.cloud.utils.script.Script.findScript(String, String) return a null value;
* com.cloud.hypervisor.xenserver.resource.CitrixResourceBaseTest.testGetPathFilesListReturned(CitrixResourceBase), this method tests the correct return value from getPatchFilePath() method, basically, verify if the returned list contain the file with the same absolute path that was retrieved from the findScript method.
We also changed the name of those test classes, the change was basically remove the Path word from the name of classes.
* pr/944:
created tests cases for method "citrixResourceBase.getPatchFiles"
Signed-off-by: Remi Bergsma <github@remi.nl>
The PlannerBase is an empty class, and does not bring contribution to
ACS code.
We changed com.cloud.deploy.FirstFitPlanner to extend the AdapterBase
and implements the com.cloud.deploy.DeploymentPlanner.
We also removed the method
com.cloud.deploy.DeploymentPlanner.check(VirtualMachineProfile,
DeploymentPlan, DeployDestination, ExcludeList) that was not used
anywhere.
Additionally, we removed the “_” from FirstFitPlanner's attributes name,
in order to have them in a more standard way.
* 4.6:
Use version for RC branch name instead of branch
make sure all files are updates with new version
Update L10N resource files with 4.6 strings from Transifex (20151129)
Fix secondary storage not working with swift
CLOUDSTACK-9083: Add disk serial to kvm virt xml
CLOUDSTACK-9081 Migration of vm across clusters fails in clustered MS setupStorage motion of vm across clusters/xenserver-pools fails in a
clustered management server setup. In xen storage motion we have to send
a migrate_receive command to the destination host, followed by
migrate_send command to the source host. The sr and network detials of
the destination host have to be passed in migrate_send command on the
source. While migrating across clusters the source and destination
resources are seperate objects. To pass this information across
resources we have to send seperate migrate with storage receive and send
commands to the resource. In a clustered ms setup these commands may
have to be forwarded to another ms as the resource may be owned by it.
The serilization of the command and answer objects fails in such case as
it doesn't understand the xapi sr and network objects. Made a change to
serialize these objects in the resource layer and pass the around as
strings in the command and answer objects.
Reviewed-By: @likitha
Author: @devdeep
Tested-By: @sailajamada
* pr/1118:
CLOUDSTACK-9081 Migration of vm across clusters fails in clustered MS setup
Signed-off-by: Remi Bergsma <github@remi.nl>
we also changed the name of those test classes, the change was basically
remove the “Path” word from the name of classes.
change the private method
com.cloud.hypervisor.xenserver.resource.CitrixResourceBase.getPatchFiles
to protected
added new class in test cases
The main changes are:
• Changing methods “Void” to “void”.
• Removal of the method “Void copyAsync(DataObject srcData, DataObject
destData, AsyncCompletionCallback callback)” that was never used. We
noticed that methods form that class are using the return type Void with
capital V. This way that method has to return a null value at the end.
commit #7
So far only 1 controller (scsi or ide) is supported in Cloudstack for ide or
scsi, this is existing limitation. Added support for 2nd IDE controller. Support adding IDE
virtual disk to VM. Also added check if VM is running as IDE virtual disk cannot be attached
to VM if VM is runnning.If user detaches a virtual disk on lower unit number of controller,
then subsequent attach operation should find free unit number on the controller and attach
the virtual disk there.
commit #6
Let the controllers of existing VMs continue without flip, current busInfo retrieved from
chain_info field of volume record from database would be preferred over
controller settings from all configuration settings.
commit #5
Editing global configuration param vmware.root.disk.controller osdefault value results
in loss of previous root disk controller type. Hence root disk's controller type for legacy
VMs is unknow post that modificaiton by user. If VM is stop/start then we could get this
infromation from bus info of existing volume. But if user resets VM and then try to start VM.
The existing bus info would be lost. Hence existing disk info is not available to depend on.
Using lsilogic or generic scsi controller for ROOT disk of legacy VMs if reset.
commit #4
Avoid adding additional (>1) scsi controllers to system vms. While attaching volume to legacy VM
don't use osdefault optoin which applicable only for VM created with the option enabled, use
legacy data disk controller type (lsilogic)
commit #3
If root disk's controller type is scsi and data disk controller type condenses
to any of scsi sub-types then data disk controller type would fall back to root disk controller itself. This
ensures data volumes would be accessible in all cases as controller of root volume would be reliable
and it means VM has the supported controller. It also avoids mix of scsi controller sub-types in a user instance.
Also translating disk controller type scsi to lsilogic.
commit #2
Support auto detection of recommended virtual disk controller type for specific guest OS.
commit #1
Support granual controller types. Add support for controller types in template registration as well.
Fix white spaces.
Removed stale HEAD merge lines
Removed tail of merge lines
Fixed VmwareResource, removing storage commands that moved to VmwareStorageProcessor.
removed stale code of controller that is present in processor
Fixed check style errors.
Fixed injection.
Tested with Linux and windows templates. Unable to run iso based tests due to few bugs in register iso area.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
(cherry picked from commit a4cc987a6f)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Storage motion of vm across clusters/xenserver-pools fails in a
clustered management server setup. In xen storage motion we have to send
a migrate_receive command to the destination host, followed by
migrate_send command to the source host. The sr and network detials of
the destination host have to be passed in migrate_send command on the
source. While migrating across clusters the source and destination
resources are seperate objects. To pass this information across
resources we have to send seperate migrate with storage receive and send
commands to the resource. In a clustered ms setup these commands may
have to be forwarded to another ms as the resource may be owned by it.
The serilization of the command and answer objects fails in such case as
it doesn't understand the xapi sr and network objects. Made a change to
serialize these objects in the resource layer and pass the around as
strings in the command and answer objects. Reviewed-By: Likitha
Adds disk serial ids based on volume uuids to the virt xml. This may be useful
for appliances/software that needs some serial ids on the VM disks. This does not
impact existing/running VMs, the vm virt xmls will be updates for running VMs
the next time they are stopped/started.
For testing, disk serial (of debian based systemvm) in the virt xml matched that
in /sys/devices/pci0000:00:0000:00:07.0/virtio4/block/vda/serial.
We currently don't support scsi-blcok devices for which serial is not supported,
for this we've added a DeviceType (LUN) which may be used in future and a check
to not add the serial to the xml if disk type is LUN.
Refer: https://libvirt.org/formatdomain.html#elementsDisks
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
CLOUDSTACK-9062: Improve S3 implementation.The S3 implementation is far from finished, this commit focuses on the bases.
- Upgrade AWS SDK to latest version.
- Rewrite S3 Template downloader.
- Rewrite S3Utils utility class.
- Improve addImageStoreS3 API command.
- Split various classes for convenience.
- Various minor improvements and code optimizations.
A side effect of the new AWS SDK is that it, by default, uses the V4 signature. Therefore I added an option to specify the Signer, so it stays compatible with previous versions.
Please review thoroughly, both code inspection and (automated) integration tests. Currently no integration tests are available specifically for S3. Therefore the implementation is needed to be tested manually, for now...
What I tested:
- Greenfield install -> will download latest systemvm template automatically to S3.
- Upload a template/iso
- Download a template/iso
- Restart of management server -> list available templates -> doesn't download them again if available.
* pr/1083:
CLOUDSTACK-9062: Improve S3 implementation.
Signed-off-by: Remi Bergsma <github@remi.nl>
CLOUDSTACK-9067 - As I developer I want to remove all the unused router-shell scripts from ACSThis PR removes the unused shell scripts that were present in the ACS project. Those script were replaced by the.
Some of the scripts are used by the HyperV Resource, which were hardcoded. I took the opportunity to use the Java constants over there as well, so the next one touching the code will know they exist and won't hardcode anything.
The following task were applied:
* Remove the shell files and the Java constants that were mapping them;
* Apply the use of the Java constants to the HyperV Resource class;
* Wrap the String.format() method in the StringUtils so we can test the changes in the HyperV Resource class.
The last point was added because I do not have a HyperV test environment. Hence, I wanted to make sure the tiny code I changed is covered at least by unit tests.
* pr/1084:
CLOUDSTACK-9067 - Replaces hardcoded paths with the VRScripts constants.
CLOUDSTACK-9067 - Fomatting the code of HypervDirectConnectResource class
CLOUDSTACK-9067 - Remove old script file from the project
Signed-off-by: Remi Bergsma <github@remi.nl>
The S3 implementation is far from finished, this commit focusses on the bases.
- Upgrade AWS SDK to latest version.
- Rewrite S3 Template downloader.
- Rewrite S3Utils utility class.
- Improve addImageStoreS3 API command.
- Split various classes for convenience.
- Various minor improvements and code optimalisations.
A side effect of the new AWS SDK is that it, by default, uses the V4 signature. Therefore I added an option to specify the Signer, so it stays compatible with previous versions.
Add Unit Tests for Libvirt/KVM storage codeThese classes were not covered by Unit Tests and this commit
adds some tests for their basic functionality.
* pr/986:
Add Unit Tests for Libvirt/KVM storage code
Signed-off-by: Remi Bergsma <github@remi.nl>
CLOUDSTACK-8964: Ovm3HypervisorGuru answer fix and snapshot to volume implementation.some clean up of code after the fix from #975
* pr/1015:
CLOUDSTACK-8964 side effect isolation extract side effect away to emphasize the main commandDelegation objective
host delegation logging for XenServerGuru.java
Getting volume from snapshot working again... odd
simple change to prevent failure and keep OVM3 snapshots working
This closes#975
Signed-off-by: Remi Bergsma <github@remi.nl>
These were lacking, but this helper is used in various places
inside the KVM code.
Some simple tests to verify the helper is doing what we expect it
to do.
CLOUDSTACK-8826: XenServer - Use device id passed as part of attach volume API properly
If device id passed as part of API and available then use it otherwise fallback on XS to automatically assign one.
For ISO device id used is 3 and it is processed before any other entry to avoid conflict.
Signed-off-by: Koushik Das <koushik@apache.org>
If device id passed as part of API and available then use it otherwise fallback on XS to automatically assign one.
For ISO device id used is 3 and it is processed before any other entry to avoid conflict.
Replaced all occurences of Charset.forName(UTF-8) with StringUtils.getPreferredCharset().
* pr/825:
Replaced all occurences of Charset.forName(UTF-8) with StringUtils.getPreferredCharset().
Signed-off-by: Daan Hoogland <daan@onecht.net>
sysctl: don't modify /etc/sysctl.confTo configure firewall rules, CloudStack modifies `/etc/sysctl.conf` and
execute those modifications. This may be harmful for several reasons:
1. `/etc/sysctl.conf` may be managed by some configuration management
system. Such a system will constantly restore the previous version.
2. `/etc/sysctl.conf` may contain additional properties that have been
changed later by some system administrator (for example, once a
firewall has been configured, forwarding may have been activated
while it is disabled in `/etc/sysctl.conf`). Executing the file
again at a later time may disrupt the system.
3. Entries are added again and again. `/etc/sysctl.conf` will contain
the same directives repeated several times.
Using a configuration file is not needed as `sysctl` is able to directly
modify sysctl values with `-w` flag.
Signed-off-by: Vincent Bernat <Vincent.Bernat@exoscale.ch>
* pr/776:
sysctl: don't modify /etc/sysctl.conf
Signed-off-by: Wido den Hollander <wido@widodh.nl>
To configure firewall rules, CloudStack modifies `/etc/sysctl.conf` and
execute those modifications. This may be harmful for several reasons:
1. `/etc/sysctl.conf` may be managed by some configuration management
system. Such a system will constantly restore the previous version.
2. `/etc/sysctl.conf` may contain additional properties that have been
changed later by some system administrator (for example, once a
firewall has been configured, forwarding may have been activated
while it is disabled in `/etc/sysctl.conf`). Executing the file
again at a later time may disrupt the system.
3. Entries are added again and again. `/etc/sysctl.conf` will contain
the same directives repeated several times.
Using a configuration file is not needed as `sysctl` is able to directly
modify sysctl values with `-w` flag.
Signed-off-by: Vincent Bernat <Vincent.Bernat@exoscale.ch>
Use host.reserved.ram.mb agent property to modify total system RAM
before reporting to management server.
Remove dom0ram variable and its min/max calculation. Just reserve a
default of 1GB, unless overridden by host.reserved.mem.mb property.
CLOUDSTACK-8762: Check to confirm disk activity before starting a VMImplements a VM volume/disk file activity checker that checks if QCOW2 file
has been changed before starting the VM. This is useful as a pessimistic
approach to save VMs that were running on faulty hosts that CloudStack could
try to launch on other hosts while the host was not cleanly fenced. This is
optional and available only if you enable the settings in agent.properties
file, on per-host basis.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
* pr/753:
CLOUDSTACK-8762: Check to confirm disk activity before starting a VM
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Guys, can you review it? things need to be discussed:
(1) this supports KVM/QCOW2 only. Anyone want to implement for other Hypervisor/format ?
(2) The original data volume (on primary storage) will be removed.
(3) The script uses the default timeout in libvirtComputingResource. Do we need to add one in global configuration (like copy.volume.wait or backup.snapshot.wait, create.volume.from.snapshot.wait)
(4) In scripts/storage/qcow2/managesnapshot.sh, I use "qemu-img convert -f qcow2 -O qcow2" to copy the snapshot from secondary to primary (hence there is no base image file), instead of "cp -f", this is because convert is faster than cp in my testing.
* pr/732:
CLOUDSTACK-5863: revert volume snapshot for KVM/QCOW2
Signed-off-by: Wei Zhou <w.zhou@tech.leaseweb.com>
Made following fixes in simulator
- Support for ScaleVmCommand/NetworkRulesVmSecondaryIpCommand in resource layer
- Added support for scaling up a running VM in simulator
- Fixed some method names not following convention
In order to test PR #725 using simulator some of these changes are needed.
Based on the way HV check is present in the scale VM API, had to explicitly put simulator related check to allow support. The ideal way would be to remove all these HV specific check from code and made them some configuration (by putting them in hypervisor_capabilities table in DB). But that will be a bigger effort outside the scope of this PR.
Signed-off-by: Koushik Das <koushik@apache.org>
This reverts commit cd7218e241, reversing
changes made to f5a7395cc2.
Reason for Revert:
noredist build failed with the below error:
[ERROR] Failed to execute goal org.apache.maven.plugins:maven-compiler-plugin:3.2:compile (default-compile) on project cloud-plugin-hypervisor-vmware: Compilation failure
[ERROR] /home/jenkins/acs/workspace/build-master-noredist/plugins/hypervisors/vmware/src/com/cloud/hypervisor/guru/VMwareGuru.java:[484,12] error: non-static variable logger cannot be referenced from a static context
[ERROR] -> [Help 1]
even the normal build is broken as reported by @koushik-das on dev list
http://markmail.org/message/nngimssuzkj5gpbz
Implements a VM volume/disk file activity checker that checks if QCOW2 file
has been changed before starting the VM. This is useful as a pessimistic
approach to save VMs that were running on faulty hosts that CloudStack could
try to launch on other hosts while the host was not cleanly fenced. This is
optional and available only if you enable the settings in agent.properties
file, on per-host basis.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
messages: Removed unused agent command AttachVolumeCommand and corresponding answer AttachVolumeAnswer
These types are not used anywhere in the orchestration layer.
Signed-off-by: Koushik Das <koushik@apache.org>
Default to notify only script to handle non-CLVM/QCOW cases.This relates to [CLOUDSTACK-8732](https://issues.apache.org/jira/browse/CLOUDSTACK-8732)
Before this commit the call to `getResizeScriptType` would throw an exception (earlier versions returned `null`, which was fine) - this caused the RBD case to fail. By changing the default to notify only we fix the case for any non-CLVM and non-QCOW cases, too.
This is RBD for now, but this should extend to new storage types supported by Libvirt natively in future.
This is my first attempted contribution: I can see a case for adding RBD logic to the actual getResizeScriptType call, too, but I felt that putting it `LibvirtResizeVolumeCommandWrapper.java` kept the special-casing of RBD (and comments about that) in one place.
### Caveat:
With Libvirt 1.2.2 this actually doesn't do the right thing - but it does do what the documentation *says* should be the right thing, so I'm going to test if this is a Libvirt bug which is fixed in a later version.
(To make it work I need to execute something like:
virsh blockresize --path vda --size 100G i-7-44-VM
where vda is the path as far as the *guest* is concerned, and not an `rbd/` path - which *should* work, but doesn't.)
* pr/696:
Default to notify only script to handle non-CLVM/QCOW cases.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
- Support for ScaleVmCommand/NetworkRulesVmSecondaryIpCommand in resource layer
- Added support for scaling up a running VM in simulator
- Fixed some method names not following convention
CLOUDSTACK-8443: Support CentOS 7 as KVM hypervisorThis adds support for CentOS 7 as a KVM hypervisor. As discussed in the Jira issue, the root cause of the issue was the `cloudstack-setup-agent` script that added this line to `/etc/libvirt/qemu.conf`:
`cgroup_controllers=["cpu"]`
On CentOS 6 this 'works', but since by default no cgroups are mounted, this setting is ignored by `libvirtd`. On CentOS 7 is does not work, because the `cpu` and `cpuacct` are 'co-mounted'. This simpy means you cannot use one without the order. The config line above instructs `libvirtd` to only use the `cpu` cgroup, which it can't.
There was code added to make sure the `cpu` cgroup was not co-mounted. It did this with a shell script, but after a reboot it was gone. It wasn't reliable and not even a solution at all. I went ahead and removed the code, because the co-mounted cgroups work fine. The real issue was the config line we put in `/etc/libvirt/qemu.conf` so much better to fix that instead.
The default for `libvirtd` is to use only cgroups that are mounted. When you remove the 'cgroup_controllers' line, the default kicks in. According to the `qemu.conf` file, the default is:
`cgroup_controllers = ["cpu", "devices", "memory", "blkio", "cpuacct", "net_cls"]`
Then again, if it's not mounted, it's not used. So, that's a nice way to control it.
This works just fine for both RHEL/CentOS 6 and RHEL/CentOS 7.
Along the debugging, I came across some small issues here and there which I also fixed in this PR.
If you test this, be sure to build RPMs from this branch and install those on the KVM hypervisor.
At Schuberg Philis we run KVM hypervisors on CentOS 7 for months with these fixes. Now I took the time to fix the issues in CloudStack, rather than work around them.
@bhaisaab @karuturi could you please review this? Thanks!
Please note:
The next step is to fix CLOUDSTACK-8625 (Systemd profile for KVM Agent) as currently a sysvinit script is still used. Although it works, we do not have any systemd benefits. As it is a separate issue, I'll try to address it in a separate PR.
* pr/731:
CLOUDSTACK-8443: detect CentOS 7.x as RHEL 7 alike
CLOUDSTACK-8443: mention the correct logfile
CLOUDSTACK-8443: display the right hypervisor type
CLOUDSTACK-8443: don't try to fix co-mounted cgroups
Signed-off-by: Remi Bergsma <github@remi.nl>
This setting works on CentOS 6 / RHEL 6 but does nothing, as
"cpu" cgroup is not mounted. On CentOS 7 / RHEL 7 systemd does
mount cgroups and "cpu" is co-mounted with "cpuacc". Hence, if
we specify "cpu" then this results in an error because it can
only use them both, or none.
By removing the setting, we rely on the default of qemu, which
is:
cgroup_controllers = ["cpu", "devices", "memory", "blkio", "cpuacct", "net_cls"]
Only if they are really mounted, they will be used. So, this will
work on both version 6 and 7.
The 'fix script' didn't work well, as after a reboot you'd still have qemu
throwing errors. Now we can handle the co-mountedcgroups.
Removed duplicate code in CitrixResourceBase and its subclassesRemoved unnecessary duplicated code by having the body of the getPatchFiles method only in the CitrixResourceBase superclass. Given that all of its implementations consisted of the same code except for the path which contains the necessary file for that implementation. An abstract method getPatchFilePath was created in the CitrixResourceBase superclass so that each implementation may return the path containing the specific file needed by that implementation.
Test cases were created for each implementation, simple as they may be. One assert is made to verify that the path returned by each implementation corresponds to the path that was previously specified on each getPatchFiles implementation.
* pr/700:
Removed duplicate code in CitrixResourceBase.getPatchFiles
Signed-off-by: Remi Bergsma <github@remi.nl>
This tries to avoid cleaning by a device name.
This closes#718
(cherry picked from commit 74f697a2dd)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Changed methodnames according to Nic.java refactor.
Fixed NicVO.java due to regression from Nic.java refactor.
Fixed VmWareGuru.java after Nic.java refactor.
See issue CLOUDSTACK-8736 for ongoing effort to clean up network code.
@cristofolini comments:
Removed unnecessary duplicated code by having the body of the
getPatchFiles method only in the
CitrixResourceBase superclass. Given that all of its implementations
consisted of the same code
except for the path which contains the necessary file for that
implementation. An abstract method
getPatchFilePath was created in the CitrixResourceBase superclass so
that each implementation may
return the path containing the specific file needed by that
implementation.
Test cases were created for each implementation, simple as they may be.
One assert is made to verify
that the path returned by each implementation corresponds to the path
that was previously specified
on each getPatchFiles implementation.
removed trailing whitespace
Cloudstack 8656: do away with more silently ignoring exceptions.a lot of messages added.
some restructuring for test exception assertions and try-with-resource blocks
* pr/654: (29 commits)
CLOUDSTACK-8656: more logging instead of sysout
CLOUDSTACK-8656: use catch block for validation
CLOUDSTACK-8656: class in json specified not found
CLOUDSTACK-8656: removed unused classes
CLOUDSTACK-8656: restructure of tests
CLOUDSTACK-8656: reorganise sychronized block
CLOUDSTACK-8656: restructure tests to ensure exception throwing
CLOUDSTACK-8656: validate the throwing of ServerApiException
CLOUDSTACK-8656: logging ignored exceptions
CLOUDSTACK-8656: try-w-r removes need for empty catch block
CLOUDSTACK-8656: try-w-r instead of clunckey close-except
CLOUDSTACK-8656: deal with empty SQLException catch block by try-w-r
CLOUDSTACK-8656: unnecessary close construct removed
CLOUDSTACK-8656: message about timed buffer logging
CLOUDSTACK-8656: message about invalid number from store
CLOUDSTACK-8656: move cli test tool to separate file
CLOUDSTACK-8656: exception is the rule for some tests
CLOUDSTACK-8656: network related exception logging
CLOUDSTACK-8656: reporting ignored exceptions in server
CLOUDSTACK-8656: log in case we are on a platform not supporting UTF8
...
Signed-off-by: Remi Bergsma <github@remi.nl>
Before this commit the call to `getResizeScriptType` would throw
an exception (earlier versions returned `null`, which was fine) -
this caused the RBD case to fail. By changing the default to
notify only we fix the case for any non-CLVM and non-QCOW cases.
This is RBD for now, but this should extend to new storage
types suppored by Libvirt natively in future.
When dumping XML use appropriate flags:
1, VIR_DOMAIN_XML_SECURE (dump security sensitive information too)
8, VIR_DOMAIN_XML_MIGRATABLE (dump XML suitable for migration)
Source:
https://libvirt.org/html/libvirt-libvirt-domain.html#virDomainXMLFlags
This fixes CVE 2015-3252: VNC password lost during VM migration across KVM
hosts. The issue is also seen when a VM is rebooted.
(cherry picked from commit cb2aca7516)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
* pr/547:
CLOUDSTACK-8601. VMFS storage added as local storage can be re-added as shared storage. Fail addition of a VMFS shared storage pool in case it has already been added as local storage in CS.
Signed-off-by: Mike Tutkowski <mike.tutkowski@solidfire.com>
On NFS we write a heartbeat, but without those we can not safely
fence off a host.
If we fence without knowing about a heartbeat we can cause a split-brain situation.
Signed-off-by: Wido den Hollander <wido@widodh.nl>
- Make sure the new password replaces the old one in the queue
- Updated the patch files for XenServer
- Updated the script path on LibvirtComputing class
- Adding update_host_passwd to VRScripts
- Add implementation to CitrixUpdateHostPasswordCommandWrapper
- Improve testUpdateHostPasswordCommand() unit test on CitrixRequestWrapperTest
- Adding update_host_passwd.sh script
- Adding the host IP address as an instance variable on UpdateHostPasswordCommand
- Improving the Unit Test (LibvirtComputingResourceTest) to get it covering the new code
- Make sure doUpdateHostPassword() doesn't get called if flag is set to false
- Do not update XenServer hosts if the cluster ID is not informed
This closes#559
- Adding more unit tests as suggested by Daan and Rajani.
- Mocking SshHelper and the Script classes, used by Citrix and Libvirt resources respectively.
- Changed location of the update_host_passwd script
- Updated the patch files for XenServer
- Updated the script path on LibvirtComputing class
- Removed the hostIP from the LibvirtUpdateHostPasswordCommandWrapper execute() method
- Adding update_host_passwd to VRScripts
- Add accessor method to host password on CitrixResourceBase
- Add implementation to CitrixUpdateHostPasswordCommandWrapper
- Improve testUpdateHostPasswordCommand() unit test on CitrixRequestWrapperTest
- Add line to patch files on xenserver directory
Concerning the LibVirt change:
- I forgot to assing the return of the getDefaultHypervisorScriptsDir() method to the hypervisorScriptsDir variable
- Modifying the LibvirtUpdateHostPasswordCommandWrapper in order to execute the script on the host
- Adding the script path to LibvirtComputingResource
- Adding the host IP address as an instance variable on UpdateHostPasswordCommand
- Improving the Unit Test (LibvirtComputingResourceTest) to get it covering the new code
- Make sure the doUpdateHostPassword() method returns the result from the answer, not always true
- Added the LibvirtUpdateHostPasswordCommandWrapper class, which will handle the changes in the KVM agent
That's will be further implemented in the future. The command will avoid the agent to complain about Unsupported command
- Added a test to make sure the current implementation of the comment works
- If changes in the future, it will also require changes. In that way, we make sure nobody will break it
Signed-off-by: wilderrodrigues <wrodrigues@schubergphilis.com>
This closes#527
CLOUDSTACK-3317 - DVS does not support management\storage network
Added support for Management and Storage Network traffic over VMware DVS in CloudStack deployments. Also added support for storage VLAN over dvPortGroup.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
CLOUDSTACK-3317 - DVS does not support management\storage network
Use non-zero dvport count while updating dvportgroups of system traffic.
Updated configuration compare logic to avoid update dvportgroup operation unless required.
This would help improve speed in vm/network deployment as the update calls would reduce. Also improved logging.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
Adding unit tests in class HypervisorHostHelperTest
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
Added license header to new file being added to repo/branch.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
Assertion is not used in runtime, correct way is throw and handle exception without killing app
Signed-off-by: Daan Hoogland <daan.hoogland@gmail.com>
This closes#362
VmwareResource.java:315, MS_SHOULD_BE_FINAL, Priority: High
com.cloud.hypervisor.vmware.resource.VmwareResource.s_serviceContext isn't final but should be
VmwareResource.java:331, MS_SHOULD_BE_FINAL, Priority: High
com.cloud.hypervisor.vmware.resource.VmwareResource.s_powerStatesTable isn't final but should be
Signed-off-by: Daan Hoogland <daan.hoogland@gmail.com>
This closes#363
KVM hosts which are actuall up, but if their agents are shutdown should be put
in disconnected state. This would avoid getting the VMs HA'd and other commands
such as deploying a VM will exclude that host and save us from errors.
The improvement is that, we first try to contact the KVM host itself. If it fails
we assume that it's disconnected, and then ask its KVM neighbours if they can
check its status. If all of the KVM neighbours tell us that it's Down and we're
unable to reach the KVM host, then the host is possibly down. In case any of the
KVM neighbours tell us that it's Up but we're unable to reach the KVM host then
we can be sure that the agent is offline but the host is running.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Signed-off-by: wilderrodrigues <wrodrigues@schubergphilis.com>
This closes#340
When executing the tests in an environment where Libvirt is also installed, it
caused errors.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
This closes#342
Maven was bulding the project and executing the tests just fine because it uses "-sourcepath". However,
with IDEs and javac it would fail.
Signed-off-by: wilderrodrigues <wrodrigues@schubergphilis.com>
This closes#317
Pull average Cpu util report between polling intervals instead of since boot
instead of using values since uptime
(cherry picked from commit 04176eaf17)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Conflicts:
plugins/hypervisors/kvm/src/com/cloud/hypervisor/kvm/resource/LibvirtComputingResource.java
This closes#297
Passing the file argument to the xml break for EL 7.1, the fix removes
the argument as just passing rombar='off' with its file arg to be empty string.
This closes#290
(cherry picked from commit aafa0c80b3)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
EL7 has a different output to 'free', use /proc/meminfo instead of a tool to be
more consistent across distros
(cherry picked from commit 212a05a345)
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Conflicts:
plugins/hypervisors/kvm/src/com/cloud/hypervisor/kvm/resource/LibvirtComputingResource.java
Removing real IPs from the tests because they cause a long running time for LibvirtComputingResourceTest
- In a local machine it takes 1.977s, but in a KVM test environment it's taking 257.879 sec
Fixing typo on LibvirtRequestWrapper
- Replace linbvirtCommands by libvirtCommands on LibvirtRequestWrapper
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
This closes#255
- The test was okay, but when running in an environment where a /root/.ssh/id_rsa existed, it would return true then fail
- We now mock the calls to methods that return the key paths, instead of relying in the static variables
- Adding LibvirtNetworkElementCommandWrapper and LibvirtStorageSubSystemCommandWrapper
- 2 unit tests added
- KVM hypervisor plugin with 22.2% coverage
I also refactored the StorageSubSystemCommand interface into an abstract class
- Remove the pseudo-multiple-inheritance implementation
- The StorageSubSystemCommand was an interface, not related to the Command class
and its implementation were extending the Command class anyway. The whole structure is better now.
- Addin LibvirtPvlanSetupCommandWrapper
- 6 unit tests added
- KVM hypervisor plugin with 21% coverage
From the 6 tests added, 2 were extra tests to increase the coverage of the LibvirtStopCommandWrapper
- Increased from 35% to 78.7%
- Adding LibvirtCopyVolumeCommandWrapper
Refactoring the LibvirtUtilitiesHelper
- Changing method name
Did not add any test to this commit due to the refactor mentioned abot.
Will proceed and add the tests
i# Please enter the commit message for your changes. Lines starting
- Gave it a better, more suggestive, name since I now added other methods to the class.
- It makes easier to mock objects and get a better coverage of the classes
- Adding LibvirtBackupSnapshotCommandWrapper, LibvirtCreatePrivateTemplateFromVolumeCommandWrapper and LibvirtManageSnapshotCommandWrapper
- 3 unit tests added
- KVM hypervisor plugin with 18.3% coverage
Less tests added to those classes because the code is quite complex and way too long.
The tests added are just covering the new flow, to make sure it works fine. I will come back to those classes later.
- Adding LibvirtOvsDestroyBridgeCommandWrapper, LibvirtOvsSetupBridgeCommandWrapper
- 4 unit tests added
- KVM hypervisor plugin with 13.9% coverage
More tests added to cover LibvirtPrepareForMigrationCommandWrapper
- Coverage of this wrapper broght from 37% to 90.6%
- 4 new tests added
Sateesh Chodapuneedi
subhash yedugundla
Rajani Karuturi
Rene Moser
Anshul Gangwar
nvazquez
Suresh Kumar Anaparti
Wei Zhou
Nitin Kumar Maharana
Rohit Yadav
Wido den Hollander
Rajani Karuturi
Mike Tutkowski
Rajani Karuturi
Bharat Kumar
Jayapal
rashmidixit
Sudhansu
Mowgli
Daan Hoogland
Nathan Johnson
Anshul And Priyank
Kishan Kavala
Anshul Gangwar
Syed
Murali Reddy
Abhinandan Prateek
Suresh Kumar Anaparti
Koushik Das
John Burwell
Sudharma Jain
Nemo
Rafael Weingärtner
Will Stevens
Aaron Hurt
Aaron Hurt
Sverrir A. Berg
weingartner
Maneesha.P
Abhinandan Prateek
NuxRo
Dmytro Shevchenko
Daan Hoogland
gabrascher
Remi Bergsma
Nitin Kumar Maharana
cirstofolini
Boris Schrijver
SudharmaJain
Nick Livens
Josh Harshman
pedro-martins
Sateesh Chodapuneedi
Devdeep Singh
Wilder Rodrigues
Funs Kessen
David Mabry
Satoru Nakaya
Vincent Bernat
Kevin McCormick
Rajani Karuturi
Likitha Shetty
Miguel Ferreira
Aaron Brady
Daan Hoogland
Funs
Wido den Hollander
Jayapal
Rafael da Fonseca
Sanjay Tripathi
