426e445811
bug 6904: during volume creation, we were NOT converting the size of the parent disk offering (which is in MB), to bytes. This is the size used to create the volume. I am adding the appropriate multipliers. Other flows like creating volume from snapshot etc should remain unaffected, as the parent volume created from the disk offering will be set right.
...
status 6904: resolved fixed
2010-11-03 11:13:06 -07:00
703fcaffe6
bug 5190: cosmetic changes for the final result returned back to the UI, for Will to consume
2010-11-03 11:13:05 -07:00
4fc8b7ae0b
bug 6892: do account check for disable user command
...
status 6892: resolved fixed
2010-11-03 10:31:16 -07:00
886fd4e754
More api refactoring - move response building for LoadBalancer/IpAddress commands to ApiResponseHelper.
2010-11-03 10:11:59 -07:00
ebdc1f069d
1) Return embedded Host object in PrepareForMaintenance command.
...
2) Build HostResonse object in ApiResponseHelper for add/reconnect/update/prepareForMaintenance/cancelMaintenance commands
2010-11-02 20:30:41 -07:00
abb8f45667
1) Return embedded object for upgradeRouter command
...
2) Build response objects in ApiResponseHelper for stop/start/reboot/list router/systemVm commands
2010-11-02 19:47:14 -07:00
694a8231ef
Build responses in ApiResponseHelper for the deploy/start/stop/reboot/recover/listVm create/list snapshot/snapshotPolicies
2010-11-02 18:16:45 -07:00
c6809f89d4
bug 6890: preventing users from searching for system account details via the list api
...
status 6890: resolved fixed
2010-11-02 17:59:24 -07:00
3c6b3b3e3d
Add more volume info to BackupSnapshotCommand, to generalize multi-hypervisor usage
2010-11-02 17:04:07 -07:00
473c3d33c3
bug 6870: during updating of pf rules, it is possible one might try to update a non existing rule via the api cmd. hitherto, we were returning a null which was assumed as a success by the cmd api, which failed with a npe. instead, we ought to throw back an error since we did not find any rule to be updated. i am checking in the same change. this also will fix the npe.
...
status 6870: resolved fixed
2010-11-02 16:07:15 -07:00
571ca42c44
1) Return embedded object for addCfg/updateCfg/listCfg.
...
2) Fixed NPE in updateCfg command (used to happen when no value was specified)
3) Fixed addCfg command to call correct getName method while setting config name (used to call the method returning the command name)
2010-11-02 15:26:33 -07:00
5fe5450abc
even more refactoring
2010-11-02 15:25:07 -07:00
d4b33f523d
bug 5190: changes from the ui perspective; we upload the entire cert as a string, instead of reading it from a file. Then, we process the cert via ip streams instead of fileinputstreams. Moved the cert validation to before cert persistance, and also added more logging and exception catch blocks to deal with IO exceptions that might arise out of the ip stream processing for the new change.
2010-11-02 14:48:05 -07:00
16438f3470
we should not be ignoring case whilst updating account, reverting to .equals instead of .equalsIgnoreCase
2010-11-02 13:45:10 -07:00
7354116517
bug 6888: there were multiple issues here. firstly, we were checking against null newaccountnames, whereas that field is a reqd val. Also, we were allowing the same name to exist across multiple accounts in the same domain. I have fixed both these issues.
...
status 6888: resolved fixed
2010-11-02 13:09:37 -07:00
560d7a275e
Removed InternalErrorException. Renamed the lock methods on the DAO objects to be more understandable.
2010-11-02 12:25:08 -07:00
3279c26796
Removed a bunch of InternalErrorException
2010-11-02 12:25:08 -07:00
6311644e08
Successful creation of domr and uservm with all programming done correctly.....yahooo
2010-11-02 12:25:08 -07:00
0c044a10d1
Moved response building for ServiceOffering/ResourceLimit commands to ApiResponseHelper class.
2010-11-02 11:59:10 -07:00
b6aeb6746c
bug 6889: fixing the issue where we were able to create a disk offering of size 0. The check was incorrect, and now we ensure we only create the right sized volumes.
...
status 6889: resolved fixed
2010-11-02 11:30:43 -07:00
85f78d8cb3
bug 6886: fixing the npe which occurs when we try to delete a non existant user. We were trying to return user details for an async job desc assuming the user always existed. Changing that logic to ensure when we don't have such details, we log the apt message
...
status 6886: resolved fixed
2010-11-02 11:20:52 -07:00
6c3256d1e4
1) Return embedded objects for enable/disable/lock/update Account/User and updateDomain
...
2) Added "haschild" return parameter to Domain response.
3) Fixed updateDomain command - used to set domainname to "updatedomainnameresponse" value if no name was specified in the request. Added more logging for updateDomain command.
2010-11-01 23:49:57 -07:00
5d4e4be974
1) Created ApiResponseHelper class - all api response objects should be built there.
...
2) Removed isMirrored tag from the update/create/listDiskOffering commands
2010-11-01 23:49:57 -07:00
3bab616478
After discussing with Alex, it was agreed that there is no need to mgmt server level locking using the mgmt_server_id. This makes the code easier, as we have one parent acquire and release mechanism, and the complexity is reduced. In accordance, we do not need to maintain cleanup tasks for mgmt server crashes, as the async mechanism deals with this. Also, the schema becomes a tad simpler as the mgmt_server_id column is not needed anymore
2010-11-01 17:04:28 -07:00
d9ba39bd8b
enable user to config default hypervisor type as vmware
2010-11-01 15:25:32 -07:00
394af1b29c
more changes on the way
2010-11-01 15:01:28 -07:00
cd1c903260
Better integration between it and virtual machine gurus
2010-11-01 15:01:28 -07:00
4db0b7d156
Better integration between it and virtual machine gurus
2010-11-01 15:01:28 -07:00
8e9d74c7f6
bug 5190: This covers the case of unforseen exceptions (although a corner case), which might pop up. We introduce a finally block which will release the cert db record for other ms to process, in case the owning ms errors out (not crash), whilst running the cert update process
2010-11-01 14:36:13 -07:00
45f4499c14
bug 6836: fixing the api responses for disk offerings. We were using the wrong objext name, and were only returning success responses. Now, we are constructing the apt object and returning the entire object back.
...
status 6836: resolved fixed
2010-11-01 13:24:28 -07:00
76c7b54fc5
fix problems after merging in bug fixes from 2.1.x
2010-11-01 12:27:15 -07:00
40f15106be
when user vm migrates, update rules on destination host
2010-11-01 12:21:56 -07:00
f70b55daa2
bug 6380: removing the code which returns hostid and hostname, as on a stopped vm, we do not have any association of the vm to the host
2010-11-01 12:16:53 -07:00
fe46ddfd63
bug 6806: changing the 'iso' to 'ISO' in all the error messages being thrown back from the API
...
status 6806: resolved fixed
2010-11-01 11:12:03 -07:00
bee976b5f5
bug 6779: returning the deviceId wherever possible, in the xxxVolume() calls
...
status 6779: resolved fixed
2010-11-01 11:12:03 -07:00
ff91abbb86
fix build break
2010-11-01 11:05:22 -07:00
1e0eb04442
Add VPN apis, some systemvm changes for vpn
2010-11-01 10:27:49 -07:00
71f37ff1c7
This is another improvement over the existing code for custom certs. We should ideally validate the cert for being a valid cert file (X.509 spec), instead of persisting it to the db and then validating it whilst constructing the ssl context. I am adding some validation around the same.
2010-11-01 10:07:53 -07:00
9b684443bb
bug 5190: Final commit to clean up an unused import, and to merge the changes with the master branch.
...
status 5190: resolved fixed
2010-11-01 09:09:18 -07:00
c74fd77730
putting in final comments, so that the code to be checked in is better readable for others trying to understand it
2010-11-01 09:06:32 -07:00
fbe2d288d7
adding some code comments
2010-11-01 09:06:32 -07:00
b9cd3e2522
some more fine tuning around corner cases
2010-11-01 09:06:32 -07:00
f3b4dd8a1c
further code modularization, extracting redundant code to common methods
2010-11-01 09:06:32 -07:00
362e67d724
adding failover protection case for when the mgmt svr crashes in the middle of executing something; along with custom response for cert
2010-11-01 09:04:51 -07:00
7b87aad660
fixed a bug in the listener flow; now handling the application of certs to existing cpvms and new cpvms based on a fork in the logic in the console proxy resource
2010-11-01 09:04:51 -07:00
abd64eddd1
adding more exception handling and debugging support
2010-11-01 09:04:51 -07:00
3752a4eb31
making the logic more robust, testing more corner cases, improving efficiency and changing the schema
2010-11-01 09:04:51 -07:00
9c80911e52
more changes
2010-11-01 09:00:58 -07:00
d9ebb7147c
improving efficiency by reducing the number of db lookups, whilst updating the custom certificates across multiple console proxies when the command is issued. Now, we use in memory hashmaps for hosts to get the hostids for cphosts, which eliminates the need to make a db lookup for every cpvm that needs to be rebooted
2010-11-01 09:00:58 -07:00
4d2e126ea1
more cpvm refactoring
2010-11-01 09:00:58 -07:00
b08037d22c
bug 6851: use specified account properties instead of admin, resourceType enum instead of type
...
status 6851: resolved fixed
2010-11-01 19:19:33 +05:30
3a0f30a049
bug 6408: remove security groups/port forwarding service in virtual network case
...
status 6408: resolved fixed
2010-11-01 14:28:03 +05:30
1b88b0eae4
bug 6748: Correcting secondary storage space utilization on the dashboard. The fix will update the existing db entry and will create a new entry for the first time only.
...
status 6748: resolved fixed
2010-11-01 12:16:07 +05:30
5c88bb936b
Oops, == doesn't work...
2010-10-29 19:51:24 -07:00
d0c68cd4a7
add a timeout (5 minutes) to wait for KVM agent host connecting to mgt svr, during adding host->discover
2010-10-29 19:18:07 -07:00
24551eea3a
Add constants for API command parameter names and for API response parameter names. Update the commands with the new constants rather than custom strings. Update several of the responses with the constants. Not all the responses have been updated, this is still work in progress...
2010-10-29 17:20:40 -07:00
f526a92b08
Only allow to attach a volume to a vm, when they are created by the same hypervisor.
2010-10-29 15:48:26 -07:00
a9349cb815
bug 6857: if the dispatcher recieves an IllegalArgumentException, bubble that up to the response so the caller gets a useful message
2010-10-29 14:07:23 -07:00
519633cd38
bug 6763: restore the volume lock when creating a snapshot for the volume to ensure only one thread is creating the snapshot at a time.
2010-10-29 13:53:00 -07:00
88dac13bb0
use the API response date format when serializing response objects
2010-10-29 13:23:29 -07:00
9cc5aaa978
Fixed listTemplates to use the template type.
2010-10-29 13:17:32 -07:00
96a4aa6f2b
checkin to allow regular code path
2010-10-29 12:16:58 -07:00
14f80f2e13
domain router up and running...now on to patching through the network between the domain router and the vm
2010-10-29 12:16:58 -07:00
b00a9174e2
always add master first when adding host
2010-10-29 11:07:51 -07:00
ad3c18a46c
bug 6789: release router lock after commiting transaction
...
status 6789: resolved fixed
2010-10-29 16:22:19 +05:30
ebac4539d0
it's possible createSnapshotDB will return null, handle that case. However, if createSnapshotDB returns null because there was no work to do (empty snapshot) we should probably throw an exception that reports that fact instead of returning null and throwing an internal error
2010-10-28 17:52:37 -07:00
642b8a56f4
If the create command returns null for whatever reason, throw an exception so it aborts scheduling the async job to act on the created object
2010-10-28 17:52:37 -07:00
640a3dc239
bug 6818: if the userAccount is null, the user is either removed or does not exist. Since we can't log them out in those cases, just skip saving the logout event (and hence avoid the NPE).
...
status 6818: resolved fixed
2010-10-28 17:52:37 -07:00
2f9e91a9ae
bug 6738: honor groupid when listing vms (restrict the list to just that group)
...
status 6738: resolved fixed
2010-10-28 16:27:07 -07:00
50a72d7f71
bug 6740: update the group for a VM if a group name is given
...
status 6740: resolved fixed
2010-10-28 16:27:07 -07:00
6435991086
bug 6782: somehow in my testing I missed this exception when creating the event for USER_DELETE. The change is to get the userAccount by 'idIncludingRemoved' just like getting the account. The proper userAccount object is returned and the event is successfully created.
...
status 6782: resolved fixed
2010-10-28 14:32:13 -07:00
7531525252
Add commented code out for the old style (2.1.x) JSON async job responses. The XML responses are still a TODO. In order to get this to fully revert back to 2.1.x style responses, some commands will need to be modified to set the responseName to be something like getResultObjectName(), and there's an example of this in StopVMCmd.
2010-10-28 14:32:13 -07:00
9eb851db60
bug 6809, 6712: standardize the domain router response to use guestmacaddress instead of macaddress
...
status 6809, 6712: resolved fixed
2010-10-28 14:32:13 -07:00
c5fef26f11
ListTemplates will no longer return any SystemVM templates. I am using a startsWith() string match for now since we used to make this comparision by ID with a single hypervisor.
2010-10-28 13:28:33 -07:00
c9069303ce
Code clean up for Extract functionality and making it more robust.
2010-10-28 18:45:12 +05:30
f0a4449ec3
ListTemplates now return the hypervisor type.
2010-10-27 19:50:36 -07:00
50c175d6c1
bug 6799: forgot to add the exception response in my commit for this bug...
2010-10-27 18:23:23 -07:00
484f334d7a
bug 6799: if there's an exception executing the Async command, create an error response with the appropriate information so that queryAsyncJobResult returns the relevant information
...
status 6799: resolved fixed
2010-10-27 18:23:23 -07:00
901f52eb99
bug 6782: incremental checkin for deleting an account. Since there was a change to findById to not find removed objects, these NPEs pop up from time to time where we expected to find the object previously and now it's not found. Since the account is getting deleted, do a findByIdIncludingRemoved to make sure the actual account object is retrieved.
2010-10-27 18:23:23 -07:00
9056653e17
Change the API Implementation annotation to use the manager's class rather than an enum. This allows methods for new managers to be added trivially rather than having to modify the dispatcher to load a new manager class.
2010-10-27 12:23:31 -07:00
a29b897b7f
bug 1974: set memory size from service offering before starting router
...
status 1974: resolved fixed
2010-10-27 12:00:11 -07:00
ae9b6d4fe1
bug 6776: for some reason, we checked for zoneId = -1 in registerTemplate and applied it to all zones, but this wasn't done for registerIso. Oh well, make the change for registerIso and things seem happy again
...
status 6776: resolved fixed
2010-10-26 17:55:00 -07:00
c9d9a18ace
bug 6746: use loadbalancer name rather than the command's name for updating the load balancer; use the load balancer's account id rather than the load balancer id when searching for the load balancer owner
...
status 6746: resolved fixed
2010-10-26 17:40:29 -07:00
4f9f249218
make console proxy work again with the new network configuration
2010-10-26 17:37:29 -07:00
e63d17dd41
bug 6739: stopSystemVm needs to return the appropriate VM instance type. The manager methods were returing VMInstanceVO, which is not sufficient for getting, for example, active viewer sessions from the console proxy object. Make sure the right VO object is returned, and all the data should be available for the API to generate a full response.
...
status 6739: resolved fixed
2010-10-26 17:33:53 -07:00
a788d2cb2e
bug 6736: rebootSystemVm should return the vm properties, not success true/false.
...
status 6736: resolved fixed
2010-10-26 17:19:55 -07:00
95da86ef1a
bug 6733: re-throw ServerApiExceptions when dispatching API commands rather than logging them and throwing a generic INTERNAL_ERROR exception
...
status 6733: resolved fixed
2010-10-26 16:54:42 -07:00
28bcbc3bdd
bug 6732: there is no longer a manual policy, so add in special handling for deleting manual snapshots. Also, in the process of deleting the snapshot, the snapshot might need to be retrieved from the database, so be sure to find including removed.
...
status 6732: resolved fixed
2010-10-26 16:47:21 -07:00
8fafc892d3
bug 6730: if a volume was never attached to a VM, then the poolId will be null and this will be flagged as not on shared storage. Since the code was deleting detached volumes only on shared storage, the never-attached volume was in an undeletable state. Instead, allow deletion of any detached volume, regardless of storage pool until some technical reason limits this from happening.
...
Also, the UI was not allowing volumes to be deleted due to a bug in looking at the vmname property of the volume. The 'if' conditional always returned true that the volume was attached to a VM to due a boolean logic error.
status 6730: resolved fixed
2010-10-26 14:56:44 -07:00
a698c0327d
adding more logging around successful cert upload
2010-10-26 14:27:26 -07:00
c6588c0e25
adding some error return handling to api dispatcher, and cleaning up some of the code paths to return the right error messages
2010-10-26 14:27:26 -07:00
42f1ff92a9
resolved merge conflicts after git pull --rebase
2010-10-26 14:27:26 -07:00
f64abad8f5
completed the listener logic. Based on the startup command, we will try and update a cust cert to a new running cpvm, and fail if such a cert doesn't exist in the cert table in our db.
2010-10-26 14:27:26 -07:00
cb4a80b8ac
starting off with the listener logic.This will ensure that every new console proxy vm that starts, will potentially look for any custom certs uploaded to our db. If they exist, we will send a command to the agent, so that the cert is uploaded, and the ssl context is re-initialized.
2010-10-26 14:27:26 -07:00
822ee2d899
more changes for the console proxy custom certificate part; some code cleanup as well as event generation with console proxy reboot logic
2010-10-26 14:27:26 -07:00
afd79a9b1b
Code cleanup
2010-10-26 13:59:11 -07:00
1b014ec340
master oss management server can start now
2010-10-26 13:47:43 -07:00
2bc938c3f8
checkins for problems.
2010-10-26 09:35:24 -07:00
29a5b796d4
deploy vm completed...now for testing
2010-10-26 09:35:24 -07:00
f58063911e
deploy vm completed
2010-10-26 09:35:24 -07:00
abhishek
alena
Kelven Yang
Alex Huang
edison
Chiradeep Vittal
root
kishan
nit
Kris McQueen
will
anthony