-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Chuck,
On 2/15/2011 5:57 PM, Caldarale, Charles R wrote: >> From: Christopher Schultz [mailto:ch...@christopherschultz.net] >> Subject: [OT] OutOfMemoryError: GC overhead limit exceeded > >> This error occurred /once/ at 16:39 EDT and it's 17:39 EDT now > >> Is it likely that this was something acute and that my server is okay? > >> PS Old Generation >> capacity = 134217728 (128.0MB) >> used = 129355080 (123.36261749267578MB) >> free = 4862648 (4.637382507324219MB) >> 96.37704491615295% used > > The above shows that the problem persists, even though at least some > work is getting done. With multiple cores available, you may be able > to get away with a saturated heap for a while, but a server restart > is in order at your earliest convenience. But before doing that, try > running jhat or equivalent to find out what's eating up the space. It seems that jmap won't do anything (like -histo) besides the heap map without using the -F switch. When I use the -F switch, the whole JVM pauses while it does the heap walk. I killed the jmap process and everything continued. I'll have to wait until off-hours to take it's temperature and bounce the server. On the other hand, I've invoked System.gc which ran quickly and freed up a lot of memory. Here's what jmap reports this morning: Heap Configuration: MinHeapFreeRatio = 40 MaxHeapFreeRatio = 70 MaxHeapSize = 201326592 (192.0MB) NewSize = 1048576 (1.0MB) MaxNewSize = 4294901760 (4095.9375MB) OldSize = 4194304 (4.0MB) NewRatio = 2 SurvivorRatio = 8 PermSize = 16777216 (16.0MB) MaxPermSize = 67108864 (64.0MB) Heap Usage: PS Young Generation Eden Space: capacity = 54263808 (51.75MB) used = 35766400 (34.1094970703125MB) free = 18497408 (17.6405029296875MB) 65.9120716334541% used - From Space: capacity = 6094848 (5.8125MB) used = 5233120 (4.990692138671875MB) free = 861728 (0.821807861328125MB) 85.86137012768818% used To Space: capacity = 6422528 (6.125MB) used = 0 (0.0MB) free = 6422528 (6.125MB) 0.0% used PS Old Generation capacity = 134217728 (128.0MB) used = 61911720 (59.043617248535156MB) free = 72306008 (68.95638275146484MB) 46.12782597541809% used PS Perm Generation capacity = 20578304 (19.625MB) used = 20152712 (19.21912384033203MB) free = 425592 (0.40587615966796875MB) 97.93184122462182% used Looks like I've got a lot more wiggle room, now, in the Old generation. Does this look healthy enough to continue, or should I schedule a restart ASAP? I'd prefer to wait until off hours. We do have a couple of operations that foolishly load way too much information into memory all at once and cache it in the session (oops), so I suspect one (or more) of those operations was performed and the objects were reachable for a while (lived in the session, probably) but then eventually became collectible and GC recovered. Quick question: does anyone know if it's safe to truncate catalina.out while Tomcat is running? Something like this: cp catalina.out catalina.out.old && > catalina.out ?? I'd like to stop getting my "the sky is falling" emails that are triggered by the presence of the string "OutOfMemoryError" in catalina.out :) Thanks, - -chris -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk1b9BcACgkQ9CaO5/Lv0PAFAACbB0afje/4WLCjgE49UnkbpWG6 OmMAoIMei7iQFVZTzZhIDcnb80W0HI8K =yPoC -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
