Tenured messgaes in GC logs after upgrading to jdk1.4.2_13

Question

our application is BEA weblogic 8.1 sp4 on solaris box,initially we were getting OOM errors but after upgrading our JVM and resizing Heap settings we are done away with OOM errors.
But in GC logs we are getting Tenured messages.
My Heap settings :
="-XX:NewSize=512m -XX:MaxNewSize=512m -XX:SurvivorRatio=4 -XX:TargetSurvivorRatio=80 -XX:PermSize=256m -XX:MaxPermSize=256m -Xms1536m -Xmx1536m
excerpts from GC logs :
261498.953: [GC 261498.954: [DefNew: 374585K->25176K(436928K), 0.2820820 secs] 1038388K->689202K(1485504K), 0.2822686 secs]
261534.897: [GC 261534.897: [DefNew: 374744K->33018K(436928K), 0.3950738 secs] 1038770K->701819K(1485504K), 0.3952532 secs]
261555.189: [GC 261555.189: [DefNew: 382586K->382586K(436928K), 0.0000436 secs]261555.189: [Tenured: 668801K->538076K(1048576K), 11.3593778 secs] 1051387K->538076K(1485504K), 11.3597008 secs]
398866.226: [GC 398866.226: [DefNew: 349568K->349568K(436928K), 0.0000439 secs]398866.227: [Tenured: 703838K->707334K(1048576K), 11.8975502 secs] 1053406K->707334K(1485504K), 11.8978785 secs]
398981.491: [GC 398981.491: [DefNew: 349568K->349568K(436928K), 0.0000434 secs]398981.491: [Tenured: 707334K->710596K(1048576K), 11.8955832 secs] 1056902K->710596K(1485504K), 11.8959308 secs]
399039.192: [GC 399039.192: [DefNew: 349568K->349568K(436928K), 0.0000456 secs]399039.192: [Tenured: 710596K->712593K(1048576K), 11.9330747 secs] 1060164K->712593K(1485504K), 11.9335147 secs]
Is there anything that we can be used to tweak our heap settings ??

Answer 1

Well, you told the JVM to use
256m permspace: -XX:PermSize=256m -XX:MaxPermSize=256m
1536m total heap: -Xms1536m -Xmx1536m
(although the GC log says 1485504K)
of which
512m new space: -XX:NewSize=512m -XX:MaxNewSize=512m
the rest is tenured space: 1024m
(here the GC log agrees saying 1048576K).
Of the 512m new space, the two survivor spaces each take 1/6 of new space giving
85m: -XX:SurvivorRatio=4
the rest of new space is eden: 340m
However the GC log says
eden: 436928K
which leaves 88m to be split between the two survivor spaces
matching -XX:SurvivorRatio=10
It looks like the GC logs totals do not include the survivor spaces as part of the heap
(not a problem as long as you know this)
and new space occupancy after GC is being mis-reported.261555.189: [GC 261555.189: [DefNew: 382586K->382586K(436928K), 0.0000436 secs]261555.189: [Tenured: 668801K->538076K(1048576K), 11.3593778 secs] 1051387K->538076K(1485504K), 11.3597008 secs]
398866.226: [GC 398866.226: [DefNew: 349568K->349568K(436928K), 0.0000439 secs]398866.227: [Tenured: 703838K->707334K(1048576K), 11.8975502 secs] 1053406K->707334K(1485504K), 11.8978785 secs]
398981.491: [GC 398981.491: [DefNew: 349568K->349568K(436928K), 0.0000434 secs]398981.491: [Tenured: 707334K->710596K(1048576K), 11.8955832 secs] 1056902K->710596K(1485504K), 11.8959308 secs]
399039.192: [GC 399039.192: [DefNew: 349568K->349568K(436928K), 0.0000456 secs]399039.192: [Tenured: 710596K->712593K(1048576K), 11.9330747 secs] 1060164K->712593K(1485504K), 11.9335147 secs]I can make sense of these numbers only if everything in new space gets promoted to tenured space on these GC runs.
Maybe you want to try running with -XX:SurvivorRatio=2