I'm seeing a configuration with one load balancer front-end running out
of heap serving six back-ends. No persistence store is configured if I
read the config correctly.  Is there a memory leak?

<server>
...
<cluster-definition id="abc">
<srun server-id="b1" host="127.0.0.1" port="6804" index="1"/>
<srun server-id="b2" host="127.0.0.1" port="6805" index="2"/>
<srun server-id="b3" host="127.0.0.1" port="6806" index="3"/>
</cluster-definition>
<cluster-definition id="a">
      <srun server-id="b1" host="127.0.0.1" port="6804" index="1"/>
</cluster-definition>
</server>

And so forth. (note: host is intentionally "localhost" in the example)

Java heap options:

/usr/jdk/instances/jdk1.6.0/bin/java -Xss1m -Xms200m -Xmx200m -Xrs

Heap after a couple days of operation:

using thread-local object allocation. 
Parallel GC with 8 thread(s) 
Heap Configuration: 
MinHeapFreeRatio = 40 
MaxHeapFreeRatio = 70 
MaxHeapSize = 209715200 (200.0MB) 
NewSize = 2228224 (2.125MB) 
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 = 24707072 (23.5625MB) 
used = 21768 (0.02075958251953125MB) 
free = 24685304 (23.54174041748047MB) 
0.08810432899535808% used 
>From Space: 
capacity = 6356992 (6.0625MB) 
used = 0 (0.0MB) 
free = 6356992 (6.0625MB) 
0.0% used 
To Space: 
capacity = 9895936 (9.4375MB) 
used = 0 (0.0MB) 
free = 9895936 (9.4375MB) 
0.0% used 
PS Old Generation 
capacity = 142606336 (136.0MB) 
used = 90085336 (85.91207122802734MB) 
free = 52521000 (50.087928771972656MB) 
63.170640608843634% used 
PS Perm Generation 
capacity = 20971520 (20.0MB) 
used = 10724648 (10.227821350097656MB) 
free = 10246872 (9.772178649902344MB) 
51.13910675048828% used

Right before failure at under one week of operation we saw:

using thread-local object allocation. 
Parallel GC with 8 thread(s) 
Heap Configuration: 
MinHeapFreeRatio = 40 
MaxHeapFreeRatio = 70 
MaxHeapSize = 209715200 (200.0MB) 
NewSize = 2228224 (2.125MB) 
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 = 14680064 (14.0MB) 
used = 14677200 (13.997268676757812MB) 
free = 2864 (0.0027313232421875MB) 
99.98049054827008% used 
>From Space: 
capacity = 11862016 (11.3125MB) 
used = 0 (0.0MB) 
free = 11862016 (11.3125MB) 
0.0% used 
To Space: 
capacity = 12451840 (11.875MB) 
used = 0 (0.0MB) 
free = 12451840 (11.875MB) 
0.0% used 
PS Old Generation 
capacity = 142606336 (136.0MB) 
used = 142606320 (135.99998474121094MB) 
free = 16 (1.52587890625E-5MB) 
99.99998878030216% used 
PS Perm Generation 
capacity = 16777216 (16.0MB) 
used = 10548616 (10.059944152832031MB) 
free = 6228600 (5.940055847167969MB) 
62.874650955200195% used

By 'failure' I mean the resin front-end is spending all it's time
garbage collecting instead of routing connections.

Thanks for any assistance.

Tom


_______________________________________________
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest

Reply via email to