Opened 12 years ago

Last modified 11 years ago

#1195 new Defect

Client crashes when restarting after Windows came out of hibernation

Reported by: Ageless Owned by: romw
Priority: Blocker Milestone: Undetermined
Component: Client - Daemon Version: 7.0.28
Keywords: Cc: romw

Description

Let me first start with saying that this thing is very difficult to reproduce. It's doing it at total random here, but having said that, it's done it 5 times already in the past 2 months.

Seen this behaviour on 7.0.25, 7.0.26, 7.0.28 and 7.0.30

When it goes crash?
Either I suspend BOINC manually (snooze from system tray menu), or I let BOINC do it itself (time-of-day) and then I either set Windows to go to hibernate, or Windows does it itself (week-days through a timer which does nothing more than C:\Windows\System32\rundll32.exe Powrprof.dll,SetSuspendState Sleep). No problem there, no the problem is when we're waking the system. When Windows comes out of hibernation, the BOINC client crashes. BOINC Manager continues to run, it's only boinc.exe that crashes.

It'll leave a message on the screen saying that BOINC has crashed and that a progress report is ready to be sent to Microsoft.

C:\Users\Ageless\AppData\Local\Temp\WERE99C.tmp.WERInternalMetadata.xml
C:\Users\Ageless\AppData\Local\Temp\WERDCF.tmp.appcompat.txt
C:\Users\Ageless\AppData\Local\Temp\WERFE2.tmp.mdmp


Exiting the error window without checking those files first results in them being dumped by Windows, you're unable to find them.

Now a new thing has come up. At the last crash BOINC was contacting Seti to download work, just as Seti had trouble with its upload & download servers. It caused 4 tasks to be stuck in perpetual download, with continual errors such as:

Can't create HTTP response output file projects/setiathome.berkeley.edu/ap_24ap12ac_B0_P0_00335_20120630_21819.wu


The last crash I had was on June 30th at 17:02:37 local time. Just 5 seconds before that, BOINC had throw errors on the Seti connect. Since that time those 4 tasks were at download stall with the above messages. Added log_snippet as attachment.

Attachments (3)

log_snippet_on_crashes.txt (3.8 KB) - added by Ageless 12 years ago.
Crashdata.zip (3.4 MB) - added by Ageless 12 years ago.
The XML, TXT and MDMDP files.
crashdumpdata.txt (4.6 KB) - added by Ageless 12 years ago.
Anyalyzation of the MiniDump? file.

Change History (11)

Changed 12 years ago by Ageless

Attachment: log_snippet_on_crashes.txt added

comment:1 Changed 12 years ago by Ageless

Summary: Client crashes when restarting after Windows is out of hibernationClient crashes when restarting after Windows came out of hibernation

comment:2 Changed 12 years ago by Ageless

Cc: romw added

Adding Rom as CC.

comment:3 Changed 12 years ago by Ageless

...Or maybe, perhaps, all it needs is for Windows to go into hibernation and come back out of it, for the strange Seti message to appear.

Just this morning:
03/07/2012 07:00:00 | | Suspending computation - time of day
03/07/2012 07:10:04 | | Windows is suspending operations
03/07/2012 07:10:05 | | Suspending network activity - requested by operating system
03/07/2012 07:10:05 | | Resuming after OS suspension
03/07/2012 10:11:12 | | Resuming network activity
03/07/2012 10:11:14 | | Windows is resuming operations
03/07/2012 10:11:14 | | [error] Can't create HTTP response output file projects/setiathome.berkeley.edu/13fe12aa.24759.7020.12.10.247
03/07/2012 10:11:14 | SETI@home | [file_xfer] project-wide xfer delay for 19737.430952 sec
03/07/2012 10:11:14 | SETI@home | Backing off 4 hr 33 min 52 sec on download of 13fe12aa.24759.7020.12.10.247
03/07/2012 10:25:47 | | [statefile] set dirty: File transfer RPC
03/07/2012 10:25:48 | | [error] Can't create HTTP response output file projects/setiathome.berkeley.edu/13fe12aa.24759.7020.12.10.253
03/07/2012 10:25:48 | SETI@home | Backing off 4 hr 5 min 1 sec on download of 13fe12aa.24759.7020.12.10.253
03/07/2012 10:25:48 | | [error] Can't create HTTP response output file projects/setiathome.berkeley.edu/12fe12ab.24140.92804.15.10.222
03/07/2012 10:25:48 | SETI@home | Backing off 3 hr 15 min 29 sec on download of 12fe12ab.24140.92804.15.10.222
03/07/2012 10:25:48 | | [error] Can't create HTTP response output file projects/setiathome.berkeley.edu/13fe12aa.24759.7020.12.10.225
03/07/2012 10:25:48 | SETI@home | [file_xfer] project-wide xfer delay for 654.717246 sec
03/07/2012 10:25:48 | SETI@home | Backing off 3 hr 22 min 26 sec on download of 13fe12aa.24759.7020.12.10.225
03/07/2012 10:26:07 | | [error] Can't create HTTP response output file projects/setiathome.berkeley.edu/13fe12aa.24759.7020.12.10.253
03/07/2012 10:26:07 | SETI@home | Backing off 3 hr 7 min 18 sec on download of 13fe12aa.24759.7020.12.10.253

I have 4 further tasks that are now stuck at downloading.

Last edited 12 years ago by Ageless (previous) (diff)

comment:4 Changed 12 years ago by Ageless

OK, almost sure now it has nothing to do with Windows hibernation, as the (7.0.31) client managed to crash in the middle of the night with the same output files. I caught them this time. They're added.

I ran the minidump file, its output is in a separate attachment. Upping the priority to Blocker.

Last edited 12 years ago by Ageless (previous) (diff)

Changed 12 years ago by Ageless

Attachment: Crashdata.zip added

The XML, TXT and MDMDP files.

Changed 12 years ago by Ageless

Attachment: crashdumpdata.txt added

Anyalyzation of the MiniDump? file.

comment:5 Changed 12 years ago by Ageless

Priority: CriticalBlocker

comment:6 Changed 11 years ago by davea

Owner: changed from davea to romw

Jord: do those files (e.g. projects/setiathome.berkeley.edu/13fe12aa.24759.7020.12.10.253) exist? If so, what's the protection on them?

Rom: why aren't we getting stack traces here?

comment:7 Changed 11 years ago by Ageless

David, that's and example of the normal Seti Multibeam task files (now Seti v7). So as long as I have work for that, I have files like that in there in the data directory. Protection? Can you explain more, what do you expect?

I do have my data directory and all subdirectories + files excluded from active scans by my antivirus and other anti-malware programs.

comment:8 Changed 11 years ago by Ageless

Already pinging another user with the crash problem. I have asked him to try with 7.2.21 and if he manages to crash that, to report here.

In the mean time, this is his thread at Seti.

Note: See TracTickets for help on using tickets.