| Summary: | Backport to 3.2.x [dstore] Hang during debug source look up | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Tools] Target Management | Reporter: | Samuel Wu <samuelwu> | ||||||||||||
| Component: | RSE | Assignee: | David McKnight <dmcknigh> | ||||||||||||
| Status: | RESOLVED FIXED | QA Contact: | Martin Oberhuber <mober.at+eclipse> | ||||||||||||
| Severity: | normal | ||||||||||||||
| Priority: | P3 | CC: | dmcknigh, kjdoyle, xuanchen | ||||||||||||
| Version: | unspecified | Flags: | kjdoyle:
review+
xuanchen: review+ |
||||||||||||
| Target Milestone: | 3.2.2+ | ||||||||||||||
| Hardware: | PC | ||||||||||||||
| OS: | Windows XP | ||||||||||||||
| Whiteboard: | |||||||||||||||
| Bug Depends on: | 358301 | ||||||||||||||
| Bug Blocks: | |||||||||||||||
| Attachments: |
|
||||||||||||||
|
Description
Samuel Wu
Created attachment 205225 [details]
backport patch
Kevin, could you please review this patch? Minor change needed for the Sender class. There is an additional "_xmlGenerator.flushData(); " on the last line changed in it that isn't required. Thanks for the review, Kevin. I took out the extra flushData() call and committed the change to the R3_2_maintenance stream. There were still uncaught memory exceptions. It hung again. The host dumps were attached. Created attachment 205445 [details]
Dump from the DSTORE server
The test was done with rseserver-M20111018-1100-linux.tar
Created attachment 205557 [details]
additional patch to deal with other out of memory cases
Xuan, could you please review the last patch? The patch looks good. Thanks. Thanks for the review, Xuan. I've committed the change to the R3_2_maintenance stream. Created attachment 206300 [details]
dstore server dumps
I picked up the driver built on 10/27 but still got the dump.
Was the fix in the 10/27 build? (In reply to comment #11) > Created attachment 206300 [details] > dstore server dumps > > I picked up the driver built on 10/27 but still got the dump. These dumps look different from previous ones. Each of these cases involve starting the FileQueryThread: at java/lang/Thread.start(Thread.java:970) at org/eclipse/rse/internal/dstore/universal/miners/filesystem /FileQueryThread.doQueryAll(FileQueryThread.java:130) at org/eclipse/rse/internal/dstore/universal/miners/filesystem/FileQueryThread.run(FileQueryThread.java:71) For this case, I could override Thread.start() to catch the OutOfMemoryError. Created attachment 206530 [details]
patch to handle additional cases
Xuan, could you review the latest patch for this? It's just a couple additional cases. The fix looks good. Thanks. Thanks for the review, Xuan. I've committed the change to the R3_2_maintenance stream. |