Commit Graph

7409 Commits

Author SHA1 Message Date
keshav 2e87fcb06d Support for inline SRX/F5. 2011-09-07 14:15:27 -07:00
Chiradeep Vittal 2c995c597b fix mocks 2011-09-07 14:07:50 -07:00
Chiradeep Vittal fbfb1b8f5b move interval to listener -- allows it to be configurable if needed 2011-09-07 13:58:49 -07:00
Chiradeep Vittal 22c1e66a6c clean up tests for security group manager v2 2011-09-07 13:58:49 -07:00
Chiradeep Vittal 3f9922ca01 Add more detail when unable to schedule cleanup 2011-09-07 13:58:49 -07:00
Chiradeep Vittal 9b75fab343 Retry only a few times 2011-09-07 13:58:49 -07:00
Chiradeep Vittal 34fe7cfc23 Don't use the same index name as before 2011-09-07 13:58:49 -07:00
Chiradeep Vittal 939946f9c6 fix a python bug 2011-09-07 13:58:49 -07:00
Chiradeep Vittal 3403ccdaf1 change some logs to debug level 2011-09-07 13:58:49 -07:00
Chiradeep Vittal 13f416698a fix merge problem 2011-09-07 13:58:49 -07:00
Chiradeep Vittal 2027049fd6 if the xenserver host cannot do bridge firewalling do not attempt to retry the security rule updat 2011-09-07 13:58:49 -07:00
Chiradeep Vittal 335feb51e4 compare numbers, not strings 2011-09-07 13:58:48 -07:00
Chiradeep Vittal cdbfb5fb0b added a few more dangerous JMX operations 2011-09-07 13:58:48 -07:00
Chiradeep Vittal 549749a2d7 Add ability to trigger ruleset updates from JMX 2011-09-07 13:58:48 -07:00
Chiradeep Vittal 65414903b7 Add the ability to disable some vms from being scheduled 2011-09-07 13:58:48 -07:00
Chiradeep Vittal 1ddda37e13 enable logging of the management server id 2011-09-07 13:58:48 -07:00
Chiradeep Vittal 78ded9c78e jconsole doesn't understand Ternary 2011-09-07 13:58:48 -07:00
Chiradeep Vittal 3c5becbd95 Enable monitoring over JMX 2011-09-07 13:58:48 -07:00
Chiradeep Vittal d531f3af29 Dequeue the vm with the largest id first -- this ensures that newly started vms get updated early 2011-09-07 13:58:48 -07:00
Chiradeep Vittal b0be029cb2 Retry 3 times on deadlock: if not updated, then throw a runtime exception to roll back the vm start 2011-09-07 13:58:48 -07:00
Chiradeep Vittal b7c1296534 make signature calculation consistent across management servers 2011-09-07 13:58:48 -07:00
Chiradeep Vittal 4b11e5e508 retry a transaction if it deadlocks 2011-09-07 13:58:48 -07:00
Chiradeep Vittal d532bec118 fix perf. issue in deploy vm. also remove dead code in dao 2011-09-07 13:58:47 -07:00
Chiradeep Vittal da857591f4 if sequence number increases but signature does not change, do not reprogram, but rewrite the rule log 2011-09-07 13:58:47 -07:00
Chiradeep Vittal c504ce1aaf if true, this could lead to the dog chasing its own tail 2011-09-07 13:58:47 -07:00
Chiradeep Vittal 8006ed159b only update sequence number when asked to. full sync SHOULD NOT request sequence number updates since this will lead to the dog chasing its own tail 2011-09-07 13:58:47 -07:00
Chiradeep Vittal ed666cb01a Avoid going back to the nics table for information we already have 2011-09-07 13:58:47 -07:00
Chiradeep Vittal 2611d1d61f thread.yield() to attempt to induce deadlocks only 2011-09-07 13:58:47 -07:00
Chiradeep Vittal 46200d122e try using multi inserts to overcome deadlock during batch update 2011-09-07 13:58:47 -07:00
Chiradeep Vittal 3c45a7a195 Reduce batch size to 16 to avoid sql errors 2011-09-07 13:58:47 -07:00
Chiradeep Vittal d6448d9fb8 useful logging 2011-09-07 13:58:47 -07:00
Chiradeep Vittal 35f44615e0 make instance_id unique to allow usage of ON DUPLICATE KEY 2011-09-07 13:58:47 -07:00
Chiradeep Vittal c8712401ca unit test for SG mgr v2 2011-09-07 13:58:47 -07:00
Chiradeep Vittal c5fea7a03c use the new implementation of SG manager 2011-09-07 13:58:47 -07:00
Chiradeep Vittal 8c6fe3a9af Use hashset since it is 50% faster. we can use treeset when we support priority 2011-09-07 13:58:47 -07:00
Chiradeep Vittal ed867d0799 unnecessary statement 2011-09-07 13:58:46 -07:00
Chiradeep Vittal 68dcad8ce6 Mocks for all 2011-09-07 13:58:46 -07:00
Chiradeep Vittal 0cc3a1e3f8 WIP: implement a queue where inserting a job for an existing vm does not create a new job 2011-09-07 13:58:46 -07:00
Chiradeep Vittal 6465ccff1b WIP : memory based queueing 2011-09-07 13:58:46 -07:00
Chiradeep Vittal fc274739f0 bug 10997: unnecessary database operation 2011-09-07 13:58:46 -07:00
Chiradeep Vittal 5172272d99 bug 10997: No need to go further when we don't have a proper job to process 2011-09-07 13:58:46 -07:00
Chiradeep Vittal eca3d1cb90 Failed waiters could wait longer than the requested timeout (upto 2x) 2011-09-07 13:58:46 -07:00
Alex Huang 3a087c2e62 schedule the work after transaction 2011-09-07 13:58:46 -07:00
Alex Huang 47957c85d9 missed one file 2011-09-07 13:58:46 -07:00
Alex Huang 0900b232fc one big transaction to see if it's any faster 2011-09-07 13:58:45 -07:00
alena 668276e22c bug 11326: don't try to transfer the agent if it's a forward agent
status 11326: resolved fixed

Also added more logging to the agent rebalance code.
2011-09-07 12:49:02 -07:00
anthony 9842a9aed3 bug 10078:
1. introduce migratewait in global configuration, the default value is 1 hour
 2. use async xapi VM migration API

status 10078: resolved fixed
2011-09-07 12:40:30 -07:00
Prasanna Santhanam ee15da14d2 setup and tests for kumquat 2011-09-07 18:34:57 +05:30
kishan 3a6c226bc9 bug 11345: Compare current and previous network stats before updating. Ignore stats if the current stats are not same as the prev stats. Set NetworkUsageAnswer log level to debug 2011-09-07 13:16:11 +05:30
alena 03958369e0 bug 11331: allow extending vnet range as a part of updateZone call
status 11331: resolved fixed
2011-09-06 17:41:38 -07:00