This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
| Summary: | [67cat] deadlock in project scan | ||
|---|---|---|---|
| Product: | serverplugins | Reporter: | nleck <nleck> |
| Component: | Infrastructure | Assignee: | Petr Hejl <phejl> |
| Status: | VERIFIED FIXED | ||
| Severity: | blocker | CC: | issues, jtulach, mmirilovic, mschovanek, pjiricka |
| Priority: | P1 | ||
| Version: | 6.x | ||
| Hardware: | All | ||
| OS: | All | ||
| Issue Type: | DEFECT | Exception Reporter: | |
| Bug Depends on: | |||
| Bug Blocks: | 161367, 162706 | ||
| Attachments: |
deadlock detected
second deadlock Screen shot of lockup. Screen shot of lockup. ( corrected mime type) |
||
|
Description
nleck
2009-05-08 02:52:06 UTC
Created attachment 81780 [details]
deadlock detected
Created attachment 81785 [details]
second deadlock
This deadlock happen twice if I tried to add a server ( which activated EE) while the project scan was running. The third time I waited for the scan to complete and then added the server and there was no deadlock. Created attachment 81788 [details]
Screen shot of lockup.
Created attachment 81789 [details]
Screen shot of lockup. ( corrected mime type)
Does not seem to be related to scanning, but rather to ergonomics&J2EE servers. it's a deadlock between them... If I want for the scan to complete ( which takes too long) there is no issue. I did it three times. I tried to set a tomcat server while waiting for the project scan to complete, each time the ide deadlocked. When I waited for the project scan to complete first the set up of the server went fine. Well, in the thread dump, the deadlock occurs during a feature enablement (ergonomics/fod). I would assume that if you wait long enough for the enablement to finish, the deadlock does not occur. Seems that waiting for the scan to finish is "long enough" for this. Petr or Jarda, so what should be the correct fix for this? Fixed in web-main 0b2fe95a8d9c. Jardo, can you review the fix? Working on unit test. Martine, should this be fixed for 6.7? I can confirm that the new patch acquires lock on ChildFactory.this while holding no other locks. This is sufficient fix of this deadlock, as it prevents #2 from http://en.wikipedia.org/wiki/Deadlock#Necessary_conditions If it's safe please fix it for 6.7. I have no objection, if it's safe - rising to P2, to keep it on RC1 Fix Candidates radar. Update to not to reschedule rescheduled update - web-main cf4310af6019. Ooops, updating priority back to P2. Integrated into 'main-golden', will be available in build *200905271401* on http://bits.netbeans.org/dev/nightly/ (upload may still be in progress) Changeset: http://hg.netbeans.org/main-golden/rev/0b2fe95a8d9c User: phejl@netbeans.org Log: #164745 [67cat] deadlock in project scan Integrated into 'main-golden', will be available in build *200905280201* on http://bits.netbeans.org/dev/nightly/ (upload may still be in progress) Changeset: http://hg.netbeans.org/main-golden/rev/cf4310af6019 User: phejl@netbeans.org Log: #164745 [67cat] deadlock in project scan (refactored to separate method) phejl, pjiricka ... are we going to fix this for NB 6.7 ? Verified in daily build by QE. Please transplant to Nb 6.7. Fixed in release clone. c0e633d336a7 b71579c29472 Verified in release67 by QE. |