[Glass] GC causing system reboot?

Mariano Martinez Peck marianopeck at gmail.com
Thu Dec 19 05:28:34 PST 2013


Hi guys,

Today morning I found out that my server was restarted automatically at
6:54 am. I was weird, since it was an unexpected restart. I tried to find
the cause and I didn't find anything from the OS itself nor from VirtualBox
logs (yes, the Linux runs over a VirtualBox machine).

Then I started to take a look to see if GemStone could have been the
cause...and indeed, there might be something related. Look at the logs (and
the timestamps):

seaside_20477admingcgem.log  shows:

--- 12/19/2013 *06:08:21*.317 EST ---
[Info]: Successfully acquired GC lockKind 6 on attempt 1.
Starting SweepWsUnion, WSU size=18  PD size=7

seaside_20475reclaimgcgem0.log shows:

12/19/2013 *06:19*:51 EST
   1 reclaims  40 pagesProcessed  40 pagesReclaimed 2 allValidPages  38
singleObjPages
   60 processedObjs  21 liveObjs  39 shadowObjs 0 deadObjs    1.5
avgObjsPerPa


All 3 fastcgi logs have several lines like:

--transcript--'handled sigabort: 2013-12-19T03:46:52.31182599067688-08:00'

Here I think the difference in the time is that the previous ones were the
server local time while in this one is the default timezone I have in
GemStone (could that be it?).

Finally, maintenance_gem.log has modification timestamp at *6:52* but no
meaningful entry. It has plenty of entries like this:

Unregistering...2013-12-19T03:52:42.86958003044128-08:00
...Expired: 0 sessions.

(again, different timestampt, but maybe the difference I mention above).


And at 6:54 (2 minutes after timestamp of mainenance_gem.log), the system
was rebooted. So....is there anything particular strange here?
I have the same setup in my machine (OSX) and I notice that every day at
some point, a topaz process starts to run in my machine and takes 100% of
CPU. It literally kills my machine, I mean, I cannot do anything else. And
the process takes quite a while. I GUESS that is the GC running. So maybe
the GC was simply trying to run in my server and the fact of being CPU
intensive put it down? mmmmm

Thanks in advance for any help,



-- 
Mariano
http://marianopeck.wordpress.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gemtalksystems.com/mailman/private/glass/attachments/20131219/b0d79cd6/attachment-0001.html>


More information about the Glass mailing list