GPU enabled hosts from non-GPU VM deployment.
Cluster reordering is based on the number of unique host tags in a cluster,
cluster with most number of unique host tags will put at the end of list.
Hosts with GPU capability will get tagged with implicit tags defined by
global config param 'implicit.host.tags' at the time os host discovery.
Also added FirstFitPlannerTest unit test file.
GPU enabled hosts from non-GPU VM deployment.
Cluster reordering is based on the number of unique host tags in a cluster,
cluster with most number of unique host tags will put at the end of list.
Hosts with GPU capability will get tagged with implicit tags defined by
global config param 'implicit.host.tags' at the time os host discovery.
Also added FirstFitPlannerTest unit test file.
Also when the ssvm is destroyed all the download urls are expired to be cleaned up in the next run by the new ssvm.
(cherry picked from commit ce90837357)
Clearly show if a volume is found and if not, that the pool is being refreshed
and the fetch is tried again.
Due to my commit b53a9dcc9f the chance of a volume
not being found is slightly bigger, but the performance gain is enormous on larger
deployments.
This is why we clearly have to log that we are refreshing the pool information
when a volume is not found.
It could be that a volume is created on host A and a few seconds later host B tries
to access the volume. In that case host B's libvirt doesn't know about the volume
yet and has to refresh the pool before it does.
(cherry picked from commit 4ee82f1f40)
2.8.7 to 3.3.5
3.3.5 is the latest stable version of AMQP client, which is also
backward comaptible. Successfully tested with updated client library.
(cherry picked from commit ff797dfa59)
modules can use it.
(cherry picked from commit b6401b04f22b0a5b686c7c477da4c6e0fd18df84)
Conflicts:
plugins/hypervisors/baremetal/src/com/cloud/baremetal/networkservice/BaremetalKickStartServiceImpl.java
GPU enabled hosts from non-GPU VM deployment.
Cluster reordering is based on the number of unique host tags in a cluster,
cluster with most number of unique host tags will put at the end of list.
Hosts with GPU capability will get tagged with implicit tags defined by
global config param 'implicit.host.tags' at the time os host discovery.
Also added FirstFitPlannerTest unit test file.
(cherry picked from commit 39fe766c2b)
Support offline volume resize on ESX by creating a worker VM to attach the unattached volume to and then resize it.
(cherry picked from commit 65ed25b7a6)
While expunging a volume, CS chooses the endpoint to perform delete operation by selecting any host that has the storage containing the volume mounted on it.
Instead, if the volume to be deleted is attached to a VM, the endpoint chosen by CCP should be the host that contains the VM.
(cherry picked from commit f1e3e83bbf)
If CCP thread local context which is used to handle connections to a vCenter is being re-used, validate that the context corresponds to the right vCenter API session.
(cherry picked from commit 6b06970366)
On larger (especially RBD) storage pools this can take a lot of
time slowing operations like creating volumes down.
The getStorageStats command will still ask a pool to be refreshed so
that the management server has accurate information about the storage pools.
On larger deployments, with thousands of volumes in one pool, this should
significantly improve storage related operations
(cherry picked from commit b53a9dcc9f)
cleaning up vmsync changes, checkvirtualmachine command was updated to return the
power state of the vm. The change was missed for Hyper-V. This caused migration to
fail on cloudstack even though it used to succeed on the hypervisor. Updated the
hyper-v agent code to return the cloudstack equivalent power state for check virtual
machine answer.
(cherry picked from commit 5350e61187)
Clearly show if a volume is found and if not, that the pool is being refreshed
and the fetch is tried again.
Due to my commit b53a9dcc9f the chance of a volume
not being found is slightly bigger, but the performance gain is enormous on larger
deployments.
This is why we clearly have to log that we are refreshing the pool information
when a volume is not found.
It could be that a volume is created on host A and a few seconds later host B tries
to access the volume. In that case host B's libvirt doesn't know about the volume
yet and has to refresh the pool before it does.
GPU enabled hosts from non-GPU VM deployment.
Cluster reordering is based on the number of unique host tags in a cluster,
cluster with most number of unique host tags will put at the end of list.
Hosts with GPU capability will get tagged with implicit tags defined by
global config param 'implicit.host.tags' at the time os host discovery.
Also added FirstFitPlannerTest unit test file.
GPU enabled hosts from non-GPU VM deployment.
Cluster reordering is based on the number of unique host tags in a cluster,
cluster with most number of unique host tags will put at the end of list.
Hosts with GPU capability will get tagged with implicit tags defined by
global config param 'implicit.host.tags' at the time os host discovery.
Also added FirstFitPlannerTest unit test file.
While expunging a volume, CS chooses the endpoint to perform delete operation by selecting any host that has the storage containing the volume mounted on it.
Instead, if the volume to be deleted is attached to a VM, the endpoint chosen by CCP should be the host that contains the VM.
If CCP thread local context which is used to handle connections to a vCenter is being re-used, validate that the context corresponds to the right vCenter API session.
On larger (especially RBD) storage pools this can take a lot of
time slowing operations like creating volumes down.
The getStorageStats command will still ask a pool to be refreshed so
that the management server has accurate information about the storage pools.
On larger deployments, with thousands of volumes in one pool, this should
significantly improve storage related operations
cleaning up vmsync changes, checkvirtualmachine command was updated to return the
power state of the vm. The change was missed for Hyper-V. This caused migration to
fail on cloudstack even though it used to succeed on the hypervisor. Updated the
hyper-v agent code to return the cloudstack equivalent power state for check virtual
machine answer.
Fix the bug by removing the command CreateEntityDownloadURLCommand from the host delegation. This results in same ssvm for creating the symlink on ssvm and same public ip being used for generating the url on MS.
NPE can happen if Spring fails to inject api authenticator, so better check
and set list of commands if the authenticator is not null or returning null cmds
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
- Return super.true() even if plugin is not enabled
- Return empty list when getCommands is called
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
Latest httpclient library suggests when creating Protocol object for use
with the apache common httpclient class, they should avoid using the
deprecated Protocol signature which takes in SecureProtocolSocketFactory
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
If session is null, probably logout (local) happened removing the name id and
session index which is needed for global logout. The limitation by design is that
local logout will void possibility of global logout. To globally logout, one
use the SLO api which would logout locally as well.
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
- Fixes signatures on plugin manager for ease of testing
- Fixes authenticator
- Adds unit testing for getType and authenticate methods for all cmd classes
- Adds SAMLAuthenticator test
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
This adds GetServiceProviderMetaDataCmd which returns SP metadata XML, since
this information should be public for IdPs to discover, we implement this as a
login/cmd api so this does not require any kind of authentication to GET this
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
- Use opensaml version from root pom
- Add utils and api as explicit dependency
- Add org.apache.cloudstack.saml.SAML2AuthServiceImpl bean
- Fix imports in all source files and resource xmls
- Use methods available from SAMLUtils to encode/decode SAML request/response
- SAML logout api is not the global logout api
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
- Move interfaces and classes from server to api module
- This can be then used for pluggable api authenticators
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
This is a feature to handle DNS entries by means of an external DNS Provider,
such as Bind. These entries include DNS domains and reverse domains, VM records
and reverse records.
For a complete description, please refer to the design document available at
https://cwiki.apache.org/confluence/display/CLOUDSTACK/Bind+and+PowerDNS+integration+by+Globo+DNSAPI
For the discussion about this feature on the dev mailing list, please refer to
http://markmail.org/thread/fvwf36hpxotiibka
Summary:
- new Network Service Provider called GloboDNS
- new Network Element to manage network domains and VM records (entries) on an external API
- new Network Resource to communicate with GloboDNS (open source)
- new API command to add DNS server
- new global option to determine if this provider should override VM entries on external DNS server
- changes in UI to include GloboDNS in Providers list
Signed-off-by: Rohit Yadav <rohit.yadav@shapeblue.com>
1. While destroying a ROOT volume do the lookup of the associated VM under the DC and not just cluster.
2. In case of VMware, during VM start if a volume is being recreated no need to detach the old volume because
we now expunge it immediately and don't wait for the storage cleanup task to run.
During VM start while configuring its disk devices, obtain the matching disk for a volume in storage
using both the volume's path and volume's datastore information.
While adding host to existing cluster which is using Nexus 1000v as a network backend, skip validation of Nexus VSM as it was already done while adding that cluster.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
determined. To exclude hosted system we filter the result on Caption='Virtual Machine'
but this string is locale dependent so it may not not work properly for locales other
than english. To overcome this now we started using ProcessId >= 0 filter
For ResizeVolume API command -
1. If hypervisor resource throws an exception, handle the NPE thrown by the job framework.
2. Improve user error message in case of RuntimeException by throwing the exception instead of 'Unexpected Exception'.
We don't need an external script to investigate the format of the RBD volume,
we only have to ask Libvirt to resize the volume and that will ask librbd to
do so.
1. Fixed JSON response deserialization. While creating a mock a JSON can be passed which will be deserialized into a response object and returned from agent layer.
For e.g. for a mock corresponding to StopCommand, a response like "{"com.cloud.agent.api.StopAnswer":{"result":false,"wait":0}}" can be passed.
2. Ability to mock PingCommand (returned as part of getCurrentStatus() agent method). As a part of this a mocked VM state report can be returned.
For e.g. {"com.cloud.agent.api.PingRoutingWithNwGroupsCommand":{"newGroupStates":{},"newStates":{},"_hostVmStateReport":{"v-2-VM":{"state":"PowerOn","host":"SimulatedAgent.e6df7732-69b2-429b-9b6a-3e24dddfa2e0"},"i-2-5-VM":{"state":"PowerOff","host":"SimulatedAgent.e6df7732-69b2-429b-9b6a-3e24dddfa2e0"}},"_gatewayAccessible":true,"_vnetAccessible":true,"hostType":"Routing","hostId":3,"contextMap":{},"wait":0}}
in the XenServer pool for GRE tunnel networks
Fix uses XenServer recommended way
Network.other_config:assume_network_is_shared=true
which ensures bridge is created automatically on hosts in the pool for
GRE tunnel networks. Fix also gets rid of error prone custom logic that ensures
bridge is created by plugging a VIF into the dom0 and connected to
GRE tunnel network.
Conflicts:
plugins/hypervisors/xenserver/src/com/cloud/hypervisor/xenserver/resource/CitrixResourceBase.java
xen cluster
XenServer does not create a bridge automatically when VIF from domU is connected
to internal network. So there is logic to force bridge creation by
creating VIF in dom0 connected to GRE tunnel network. But there is no
logic to delete the VIF after bridge gets created. So this fix ensure
VIF is delted when atleast there is one domU VIF connected to the
network.
1. Adding the missing Template/Volume URLs expiration functionality
2. Improvement - While deleting the volume during expiration use rm -rf as vmware now contains directoy
3. Improvement - Use standard Answer so that the error gets logged in case deletion of expiration link didnt work fine.
4. Improvement - In case of domain change, expire the old urls
In situations where libvirt lost the storage pool the KVM Agent will re-create the
storage pool in libvirt.
This could be then libvirt is restarted for example.
The object returned internally was missing essential information like the sourceDir
aka the Ceph pool, the monitor IPs, cephx information and such.
In this case the first operation on this newly created pool would fail. All operations
afterwards would succeed.
The proxy-arp add/del is done on firewall rule add/del.
The proxy-arp rule is deleted only when there is no static nat or dest nat rule is not using the ip.
When there is static nat or PF and firewall rule
a. Delete firewall rule. It skips delete proxy-arp because the rule is used by static nat rule.
b. After deleting fw rule if we disable static nat there is no way to delete proxy-arp rule.
On VM expunge we are deleting firewall rules first then static nat rules. This caused the stale proxy-arp
rules.
With this fix adding/deleting proxy arp rule on static nat/PF rule add/del.
to get nic info of 0.0.0.0. To get it, it iterates through all nics and return the last NIC in
the list if it doesn't match with any IP address. In case last NIC doesn't have unicastAddress,
Hyper-V agent will fail to start. We don't need IP address during initialization. It get
initialized with startupcommand later.
last VM from the VPC is deleted on a host
OVS distributed routing: ensure bridge is deleted when last VM from the
VPC is deleted on a host. This fix ensures that bridge is
destroyed.
the folder column. For an smb share the smb credentials are in the query string of the path.
Before adding the path, smb shares query string should be cleaned up.
that case cloudstack was not doing anything and not updating the state of the vms to stopped.
Now the agent returns empty list of hostvmstatereport. Management server will then update the
vm state to stopped (instead of not acting upon the return state).
not created already when OvsVpcPhysicalTopologyConfigCommand update is
recived
Currently if the tunnel creation fails, there is no retry logic. Fix
ensures OvsVpcPhysicalTopologyConfigCommand updates as an opputiunity to ensure
proper tunnels are established between the hosts.
Check if switch name detected from traffic label for management, storage, control traffic is null before falling back to default value.
Signed-off-by: Sateesh Chodapuneedi <sateesh@apache.org>
We used to create the snapshot after the copy from Secondary Storage,
but it could be that we never use the snapshot.
Now we check if the snapshot exists prior to performing the cloning operation
Sanjay Tripathi
Santhosh Edukulla
Frank Zhang
Anthony Xu
Rajesh Battala
Damodar
Nitin Mehta
Edison Su
Wido den Hollander
amoghvk
Murali Reddy
Hugo Trippaers
Likitha Shetty
Anshul Gangwar
Rohit Yadav
Devdeep Singh
David Nalley
Mike Tutkowski
Daan Hoogland
imduffy15
Kishan Kavala
Rajani Karuturi
Fred Clift
Vania Xu
Kishan Kavala
Daniel Vega
Wido den Hollander
Sateesh Chodapuneedi
Min Chen
Ritu Sabharwal
Sheng Yang
Will Stevens
Suresh Ramamurthy
Marcus Sorensen
punith-cloudbyte
Joris van Lieshout
Kelven Yang
ynojima
Vania Xu
Sujaya Maiyya
Koushik Das
Funs Kessen
Girish chaudhari
Tim Mackey
Jayapal
Harikrishna Patnala
Bharat Kumar
Sachchidanand Vaidya