[Xen-bugs] [Bug 530] memory leak eventually causes oom-killer

[bugzilla-daemon@xxxxxxxxxxxxxxxxxxx]

http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=530





------- Additional Comments From andy@xxxxxxxxxxxxxx  2006-02-20 13:05 -------
(In reply to comment #9)
> >   OBJS ACTIVE  USE OBJ SIZE  SLABS OBJ/SLAB CACHE SIZE NAME
> > 2021784 2021783  99%    0.06K  33144       61    132576K size-64
> 
> Monitor your size-64 over *time*. an instant measure doesn't tell you if 
> there's
> a leak or not there. However I never see the size-64 slab grow that much, so
> there's a good chance that's what it is.

Sure, it's just difficult to get logs of that as slabtop isn't very amenable to
redirecting to a file.  If I appended the contents of /proc/slabinfo to a file
every hour or so would this be of any use to you?

> Please confirm that your size-64 slab is almost always growing.

Anyway, the server just died again with a ream of oomkillers so here's the stats
just after a clean boot:

Mon Feb 20 13:16:48 UTC 2006, system has been up 24 minutes, with 4 guests
running one of which has been under constant heavy network load since boot (this
problem happens in dom0 regardless of whether the guests are loaded or idle).

 Active / Total Objects (% used)    : 211470 / 217844 (97.1%)
 Active / Total Slabs (% used)      : 5589 / 5591 (100.0%)
 Active / Total Caches (% used)     : 80 / 125 (64.0%)
 Active / Total Size (% used)       : 21241.07K / 22187.16K (95.7%)
 Minimum / Average / Maximum Object : 0.01K / 0.10K / 128.00K

  OBJS ACTIVE  USE OBJ SIZE  SLABS OBJ/SLAB CACHE SIZE NAME
156526 156525  99%    0.06K   2566       61     10264K size-64
  9968   9786  98%    0.14K    356       28      1424K dentry_cache
  8400   8385  99%    0.05K    112       75       448K buffer_head
  6896   6892  99%    0.47K    862        8      3448K ext3_inode_cache

I'll update this bug with the figures over time.


-- 
Configure bugmail: 
http://bugzilla.xensource.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

_______________________________________________
Xen-bugs mailing list
Xen-bugs@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-bugs