Fixed issues with vif scripts on 5.6FP1
Fixed ipv6 issue on 5.6FP1
Plus other various fixes and improvements
Starting to remove debug code
NOTE: Network is configured correctly but instances do not start. Possibly indefinite wait occuring on some commands
Fixed issues with vif scripts on 5.6FP1
Fixed ipv6 issue on 5.6FP1
Plus other various fixes and improvements
Starting to remove debug code
NOTE: Network is configured correctly but instances do not start. Possibly indefinite wait occuring on some commands
-Create multiple physical networks if there are distinct tags found in network_tags table.
-One physical network per tag.
-Currently all tags flow to xenserver traffic type label.
Description:
Fixing two other scenarios apart from the reported one
where we were not passing in database IDs for translation
into uuids, in the exception.
Only DHCP entry need to know if no one apply the entries(when VM is starting
up), other rules should be safe when return true anyway.
status 14470: resolved fixed
Changes:
- While listing UP hosts in the cluster of the host, the query has changed and now also considers the Host 'resourceState' field and picks only those hosts that have value 'enabled'.
This query result does not include the host being put in maintainance; unlike the earlier query that used to include it.
- But the logic to find out if this is the last host in the Cluster is not updated and has a bug when it compares the list size.
- Fixed this.
2) Added new api - changeServiceForSystemVm - to support service offering upgrade for system vms
3) Removed global config parameters that are not in use anymore: consoleproxy.ram.size, consoleproxy.cpu.mhz, secstorage.vm.ram.size, secstorage.vm.cpu.mhz
Reviewed-By: Sheng Yang
Changes:
Added 'removed' column to physical_network_service_providers to avoid the Foreign Key constraint error.
Conflicts:
setup/db/db/schema-30to301.sql
1) Support HTTP keep-alive in clustering communication channel
2) Increase concurrency level for clustering message delivery
Reviewed-By: Kelven (with unit test)
1) Drop synchronized call semantic for ClusterManagerImpl.broadcast()
2) Have no choice now but to use an unbound thread pool to notify upper layer. This is to prevent thread starvation when we have cross-management server waitings.
Reviewed-By: Kelven(with unit test)
It's not a elegant fix. The status for firewall rules should remain unchanged
before/after ip association/disassociation. But the related change is tricky
than this fix, may not get enough test for 3.0.1. So we would apply existed
firewall rules again, which would work, just result in some unnecessary
commands.
status 14484: resolved fixed
Reviewed-by: Edison Su
status 14500: resolved fixed
reviwed-by: Frank Zhang
Conflicts:
server/test/com/cloud/network/MockNetworkManagerImpl.java
setup/db/db/schema-30to301.sql
Changes:
Fixed as described in the bug.
* CreateVlanIpRangeCmd still accept account/domainId info
* if account owns:
- one Isolated network with source nat service enabled, use this network
- more than one Isolated network with source nat service enabled - error out
- none Isolated networks with source nat service enabled, create it only in
case when there is an Isolated network offering with Availability=Required and
source nat service enabled.
The routing table with two nics may be messed up, due to we sent same
router(gateway) information from different DHCP server, in order to specify
default gateway. E.g.
Network A: 192.168.1.0/24, gw 192.168.1.1
Network B: 192.168.2.0/24, gw 192.168.2.1
User VM: Nic 1 connect to network A, get ip 192.168.1.10; nic 2 connect to
network B, get ip 192.168.2.10.
Set network A as the default network of user VM.
Currently we would send this information to user VM through DHCP offer:
In network A: dhcp-option:router 192.168.1.1
In network B: dhcp-option:router 192.168.1.1
So both NIC in the guest VM would receive 192.168.1.1 as router(gateway).
But, in CentOS 5.6, dhclient-scripts try to tell if the gateway is reachable
for current subnet.
So when we try to enable nic 2(eth1) of user VM, dhclient would receive:
IP: 192.168.2.10
Mask: 255.255.255.0
Router: 192.168.1.1
Then it would found that the specified gateway(router) is not within its own
subnet(192.168.2.0/24). But since we send out this ip(192.168.1.1) as the
gateway for it, dhclient thought that it should got someway to access the
network through this IP. So it would execute:
ip route add 192.168.1.1 dev eth1
ip route replace default via 192.168.1.1 dev eth1
But it can never reach 192.168.1.1(which is in the eth0's subnet and the
gateway of eth0) by go through eth1 interface. So it is messed up.
We've tested Windows 2008 R2, CentOS 5.3, CentOS 5.6 and Ubuntu 10.04. Windows
and Ubuntu are fine with above policy.
To solve this, we send different dhcp:router option according to the guest OS
type now.
We may need expand this list later, but for now we only know that CentOS and
RHEL would behavior in this way.
status 14042: resolved fixed
Changes done:
- Provide UUID for userid and domainid only while constructing the login response. Session will hold the DB id's as before, to ensure other parts keep working.
- This reverts commit de28aa3ddde5b601f2f234f2eccef871fbaf1a06.
Description:
Incorrectly removed part of the XML serializer that serialized
the IdentityProxy object in normal responses, when putting in
support for serialization of lists of IdentityProxy objects in
exception responses as part of the code changes put in for bug
13217, resulting in this bug. Putting it back in place.
Description:
Adding overloaded addProxyObject() function to CloudException
and RuntimeCloudException classes and using this function
to stuff exceptions with IDs, to reduce code footprint.
Description:
Modifying the API functions' exception handling to call
addProxyObject() wherever applicable, and removing some
wrong calls to addProxyObject() that were put in in an
earlier commit for this bug.
With this commit, we cover many API functions to use the
new exception handling code, but some pieces may still be
left out. These will be covered as work in progress, when
making changes to the CS API code.
Description:
Removed some wrong invocations to addProxyObject() when
throwing exceptions in NetworkManagerImpl.java.
Replaced db ids with uuids in various points in the code
of NetworkManagerImpl.java, where exceptions are thrown.
Description:
Replacing placement of db ids in exception messages to uuids
in the file ManagementServerImpl.java.
Since there are a large number of files that throw exceptions
with db ids in them and they need to be changed, we will make
the changes in multiple commits.
Description:
Adding a new class AnnotationHelper, that provides routines
to read annotations from a VO class.
Cloudstack does the mapping between cloudstack java VO objects
and the database tables using cglib. cglib creates proxy objects
as the maps between VO objects and the database. A VO (value
object) class is populated after querying from the MYSQL database.
Ideally, a getAnnotation() issued on a cglib proxy object should
get a list of all the annotations in all classes in the inheritence
chain. However, this functionality seems to be currently broken
in cglib. Hence, when querying for annotations given the object
of a VO class, we need to get to the corresponding VO class of that
cglib proxy class and issue a getAnnotation() on that class. To get
the VO class, we simply need to get the super class of the proxy
object. Also, we need to recurse to the root VO class in case the
VO class extends another VO class.
Note that the cache used by CS is ehcache.
Bug 13127: API error text refer to database ids instead of uuids
Code-Reviewers: Ewan Mellor, Kelven Yang
Description:
1. A new class CSExceptionErrorCode has been added to utils.
It contains a list of error codes for each type of
Exception class. Use fully qualified package paths for
Exceptions in CSExceptionErrorCode. We log any exception
name not found in the list of error codes for exceptions.
2. Whenever we throw an exception exobj anywhere in the
CS code, the CSErrorCode is set in the base class
constructor.
3. We add a new field csErrorCode in classes CloudException,
RuntimeCloudException, ExecutionException and
ExceptionResponse.
4. Two places in ApiServer.java were wrongly modified when
putting in changes for bug 13127 to not throw an exception.
This has been corrected in this commit.
Bug 14006 - Admin could not create a VM when the cluster is Disabled
Changes:
- For Root admin, planner will not filter out the disabled pods or clusters from the resource list
Description:
Modified the IdentityTypeAdapter's custom serializer to
identify whether this is an exception response that is being
serialized, by checking if the idFieldName is set. If so,
serialize both uuid and the uuidProperty (for eg, zoneId and
"zoneId" (string)) and pass back the json representation of that.
Modified XML serializer also to build a list of uuids+fieldnames.
Introduced a new field "cserrorcode" in ExceptionResponse. This
refers to an error code that can be according to the specific
Exception being thrown. This will be serialized as usual. There
shouldn't be any need to do a db lookup for conversion for these
error codes.
Description:
Modify Exception handling to enable addition of multiple
uuids in a single exception thrown by API functions. Both
XML and JSON outputs will store all uuids and Fieldnames.
This will make it easier to provide more information when
an exception occurs - for example, a zone id, a cluster id,
host id, and then a specific property id.
Description:
Added a field name for the db id in the IdentityProxy class, and
modified setProxyObject() to take an additional id name parameter.
This will let us know the name of the uuid that we are returning.
E.g.- domainId, zoneId, etc. The client can view this field in
the json/xml output. Modified the JSON/XML serialization routines
to append this new parameter to the serialized output for Exception
Responses.
Description:
1) Put in an IdentityProxy object in the ExceptionResponse class.
This allows us to copy over the IdentityProxy object contained
in the exception caught by handlerequest() when thrown by the
command's execute() method, into the Response object that is
prepared to return an exception response to the calling API
invocation.
2) Modified the GSON serialization method to conver the entire
exception object into JSON format and not just the error text.
3) Modify the updateDomain API to populate the exception it throws
upon detecting a duplicate domain to include the tablename and
domain db id in the exception's IdentityProxy object.
NOTE:
1) We can modify the base exception classes and the ExceptionResponse
class to contain a list of IdentityProxy objects rather than a
single one.
2) We will need to modify all commands such that wherever applicable
(wherever a db id is involved), they populate the IdentityProxy
object(s) before throwing an exception.
Description:
1) Added a setProxyObject() method to CloudException and RuntimeCloudException
2) Modified a bunch of throw exceptions in NetworkManagerImpl.java to call setProxyObject() before throwing an exception.
3) Changed scope of ProxyIdentity attribute to protected.
4) Added routines to ServerApiException to get/set IdentityProxy object, and
routine in RuntimeCloudException to get the Idproxy object.
5) Modified the exception handling around the dispatcher and handlerequest()
to copy over the IdentityProxy information before rethrowing an exception
eventually back to handle().
6) Removed duplicate IdentityProxy object in ServerApiException.
It was extending RuntimeCloudException which already had an
IdentityProxy object.
Description:
1) Moved RuntimeCloudException from api/ to utils/.
Added simple constructor to RuntimeCloudException.
Modified all classes that extended RuntimeException
to extend RuntimeCloudException. These classes
are listed below:
ServerApiException
CloudAuthenticationException
CloudExecutionException
AsyncCommandQueued
HypervisorVersionChangedException
RuntimeCloudException
2) Added overloaded constructed to CloudException.
Modified all classes that extend Exception to extend CloudException instead.
These classes are listed below:
ConcurrentOperationException
ConflictingNetworkSettingsException
ConnectionException
DiscoveryException
InsufficientCapacityException
ManagementServerException
ResourceUnavailableException
VirtualMachineMigrationException
AgentControlChannelException
OperationTimedoutException.java
UnsupportedVersionException.java
UsageServerException.java
UnableDeleteHostException.java
AgentAuthnException.java
HttpCallException.java
ActiveFencingException.java
ClusterInvalidSessionException.java
GreTunnelException.java
OvsVlanExhaustedException.java
Description:
1) Adding two new classes, CloudException and RunTimeCloudException.
The former extends Exception and the latter RunTimeException.
These will be used by classes that formerly directly extended
Exception and RuntimeException. These two classes have an attribute
of type IdentityProxy to enable exceptions fill in db ids in separate
attribute fields rather than in a string. Doing so will allow the
serialization module (GSON for JSON and other for XML) to kick in
and convert this db id to a uuid in ApiServer.java just before the
JSON/XML responses are sent out.
2) Moving IdentityProxy.java from api/ to utils/ since
both CloudException and RuntimeCloudException refer to it.
3) Changing references to IdentityProxy class from api/ to utils/.
4) While rebasing to master, a new file was added, merging
api/src/com/cloud/api/response/IsoVmResponse.java to this diff.
Changes for Script.java:
1. Even the script is non-timeout one, set default timeout to one hour. This avoid a wrongn script forever hang
2. When InterruptedException happens, check if timeout is really reached, if not, continue
status 13962: resolved fixed
reviewed-by: Alex, Edison
Reviewed-by: Sheng Yang
Changes:
- Do not check if allocation_state is 'Enabled' in planner if the caller is Root Admin.
- This should let Root Admin create a VM in a disabled Zone.
Reviewed-by: Kishan
Changes:
- Separated out the External Network Usage task from the ExternalLBDeviceMgr because ExternalLbDeviceMgrImpl :: start() was getting multiple times during management server satrtup. The reason for this is that this is the baseclass for F5 and NetScalarElement.
- This caused us to schedule the ExternalNetworkUsageTask multiple times
- Also we have LBRulesMgr calling this ExternalLbDeviceMgrImpl by creating an instance of this class which is declared abstract
- Hence having a separate implementation to manage the network usage stats should solve this.
- configuring unique persistence profile for each LB rule with sticky method applied
- removing source based sticky method for source based LB method which is not supported by F5
And per Alex's request, add default value directly into the database, rather
than using it at last minute of implemention.
status 13829: resolved fixed
Reviewed-by: Alex
Reviewed-By: Alex
Changes:
- Reuse the same storagepool where the Volume is ready on each retry of VM deployment until the cluster where the volume is has capacity
- After the cluster is out of capacity, we look in other clusters and find a new storagepool.
- At this point if the volume is recreatable on the new storagepool, depoyment will succeed provided everytyhing else goes through
- But if the volume is not recreatable and its cluster is out of capacity, we will still fail to deploy the VM
Salvatore Orlando
Alena Prokharchyk
root
unknown
anthony
Abhi
Nitin Mehta
kishan
Kelven Yang
Sheng Yang
prachi
Rajesh Battala
Vijayendra Bhamidipati
frank
Alex Huang
David Nalley
Edison Su
Sheng Yang
Murali reddy
Likitha Shetty
U-CITRITE\vijayendra1
Edison Su