Created attachment 702848[details]
log
Description of problem:
Deadlock in DB during creation pool of vms
Version-Release number of selected component (if applicable):
Latest upstream: commit 7fd9d7045a064d5bceaf5e7be5aaceb40746bc76
How reproducible:
Steps to Reproduce:
1. Initiate creation of pool of vms (In my case pool with 15 vms)
Actual results:
Part of vm was not created because of deadlock in DB
Expected results:
No deadlock should be
Additional info:
I have one host, and 3 running vms, so this is not because of load.
Regards reproduce: it is obvious from exception how to fix it.
It is upstream bug.
Steps, like I wrote - I created a pool of vms, log attached. I success to reproduce it couple of times.
Due to logs a bug can be solved.
The deadlock is:
EntityAsyncTask::EndCommandAction [within thread]: EndAction for action type AddVm threw an exception: javax.ejb.EJBTransactionRolledbackException: CallableStatementCallback; SQL [{call updateimagestatus(?, ?)}]; ERROR: deadlock detected
Detail: Process 13637 waits for ShareLock on transaction 1692869; blocked by process 13514.
Process 13514 waits for ShareLock on transaction 1692880; blocked by process 13637.
Hint: See server log for query details.
Where: SQL statement "UPDATE images
SET imageStatus = v_status
WHERE image_guid = v_image_id"
Created attachment 702848 [details] log Description of problem: Deadlock in DB during creation pool of vms Version-Release number of selected component (if applicable): Latest upstream: commit 7fd9d7045a064d5bceaf5e7be5aaceb40746bc76 How reproducible: Steps to Reproduce: 1. Initiate creation of pool of vms (In my case pool with 15 vms) Actual results: Part of vm was not created because of deadlock in DB Expected results: No deadlock should be Additional info: I have one host, and 3 running vms, so this is not because of load. Regards reproduce: it is obvious from exception how to fix it.