We are running CFMX 6.1 on linux RHEL4.  

Approximately once every 8-12 hours or so our server stops responding for about 
3 minutes.  Symptoms: CF stops processing requests, stops queueing, stops 
timing out requests, stops doing everything.  

We are using Jrockit 1.4.2 right now, but it also happens under Sun's JVM too.

cfstat 1:

0   11  0   23  -1  -1  127   15    1     137    247    0      0      0
0   11  0   23  -1  -1  127   15    1     137    247    0      0      0
0   11  0   23  -1  -1  127   15    1     137    247    0      0      0
0   11  0   23  -1  -1  127   15    1     137    247    0      0      0
0   11  0   23  -1  -1  127   15    1     137    247    0      0      0
... etc.. for 3 minutes. 

Then it jumps back to life again, taking off as if nothing ever happened. 

I thought perhaps it was a problem with garbage collection, but Jrockit's GC 
debugging looks fine.  In fact, it continues to garbage collect during that 3 
minutes, just as it did before the freeze:

[gcpause][Thu Nov  9 16:00:07 2006][13912] old collection phase 0-2 pause time: 
7.998000 ms, (start time: 21226.183 s)
[gcpause][Thu Nov  9 16:00:07 2006][13912] total mark time: 530.209 ms
[gcpause][Thu Nov  9 16:00:07 2006][13912] total sweep time: 5.027 ms
[gcpause][Thu Nov  9 16:00:07 2006][13912] old collection phase 4-0 pause time: 
27.229000 ms, (start time: 21226.692 s)
[gcpause][Thu Nov  9 16:00:07 2006][13912] (pause includes compaction: 0.000 ms 
(no compaction), update ref: 0.000 ms)
[gcpause][Thu Nov  9 16:01:07 2006][13912] old collection phase 0-2 pause time: 
10.371000 ms, (start time: 21286.774 s)
[gcpause][Thu Nov  9 16:01:08 2006][13912] total mark time: 452.252 ms
[gcpause][Thu Nov  9 16:01:08 2006][13912] total sweep time: 39.125 ms
[gcpause][Thu Nov  9 16:01:08 2006][13912] old collection phase 4-0 pause time: 
62.351000 ms, (start time: 21287.204 s)
[gcpause][Thu Nov  9 16:01:08 2006][13912] (pause includes compaction: 30.253 
ms (external), update ref: 4.000 ms)
[gcpause][Thu Nov  9 16:02:08 2006][13912] old collection phase 0-2 pause time: 
15.459000 ms, (start time: 21347.301 s)
[gcpause][Thu Nov  9 16:02:09 2006][13912] total mark time: 581.907 ms
[gcpause][Thu Nov  9 16:02:09 2006][13912] total sweep time: 147.318 ms
[gcpause][Thu Nov  9 16:02:09 2006][13912] old collection phase 4-0 pause time: 
181.771000 ms, (start time: 21347.849 s)
[gcpause][Thu Nov  9 16:02:09 2006][13912] (pause includes compaction: 131.157 
ms (internal), update ref: 16.001 ms)
[gcpause][Thu Nov  9 16:03:09 2006][13912] old collection phase 0-2 pause time: 
16.870000 ms, (start time: 21408.057 s)

So the JVM appears to be fine.  The problem is not with the database -- there 
are no outstanding queries when this occurs.  The network connection is fine.  
The disk is fine.  There's a ton of memory.  There's no paging going on.

So... does anyone have any idea what might be causing these infrequent yet long 
pauses?

Regards,
Terry

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
Introducing the Fusion Authority Quarterly Update. 80 pages of hard-hitting,
up-to-date ColdFusion information by your peers, delivered to your door four 
times a year.
http://www.fusionauthority.com/quarterly

Archive: 
http://www.houseoffusion.com/groups/CF-Server/message.cfm/messageid:6156
Subscription: http://www.houseoffusion.com/groups/CF-Server/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.10

Reply via email to