FATAL system - out of memory.

[soltecsis]

Hello,
I have a serious problem with one mailbox that affects our Zimbra system making it crash. We have been using the system in production enviroment for several months without any problem since now.

At this moment we have the last estable version of ZCS Open Source Edition (6.0.7) over Ubuntu 8.04 64Bits. I have just updated from 6.0.6 last night but it has not solved the problem.

The problem is that every time one new message arrives to the problematic mailbox, it doesn't gets delivered and in the /opt/zimbra/log/mailbox.log appends a register like this:
[...]
2010-07-01 13:16:53,230 FATAL [LmtpServer-509] [name=user@domain.com;mid=118;ip=10.99.3.56;] system - out of memory
java.lang.OutOfMemoryError: Java heap space
at java.util.Arrays.copyOf(Arrays.java:2882)
at java.lang.AbstractStringBuilder.expandCapacity(Abs tractStringBuilder.java:100)
at java.lang.AbstractStringBuilder.append(AbstractStr ingBuilder.java:390)
at java.lang.StringBuilder.append(StringBuilder.java: 119)
at com.zimbra.cs.index.ZimbraAnalyzer$FieldTokenStrea m.addToken(ZimbraAnalyzer.java:291)
at com.zimbra.cs.index.ZimbraAnalyzer$FieldTokenStrea m.bufferNextLine(ZimbraAnalyzer.java:370)
at com.zimbra.cs.index.ZimbraAnalyzer$FieldTokenStrea m.next(ZimbraAnalyzer.java:224)
at org.apache.lucene.analysis.TokenStream.next(TokenS tream.java:91)
at org.apache.lucene.index.DocInverterPerField.proces sFields(DocInverterPerField.java:134)
at org.apache.lucene.index.DocFieldConsumersPerField. processFields(DocFieldConsumersPerField.java:36)
at org.apache.lucene.index.DocFieldProcessorPerThread .processDocument(DocFieldProcessorPerThread.java:2 34)
at org.apache.lucene.index.DocumentsWriter.updateDocu ment(DocumentsWriter.java:765)
at org.apache.lucene.index.DocumentsWriter.addDocumen t(DocumentsWriter.java:743)
at org.apache.lucene.index.IndexWriter.addDocument(In dexWriter.java:1917)
at org.apache.lucene.index.IndexWriter.addDocument(In dexWriter.java:1895)
at com.zimbra.cs.index.LuceneIndex.addDocument(Lucene Index.java:220)
at com.zimbra.cs.index.MailboxIndex.indexMailItem(Mai lboxIndex.java:513)
at com.zimbra.cs.mailbox.IndexHelper.indexingPartOfEn dTransaction(IndexHelper.java:1046)
at com.zimbra.cs.mailbox.Mailbox.endTransaction(Mailb ox.java:6815)
at com.zimbra.cs.mailbox.IndexHelper.indexItemList(In dexHelper.java:908)
at com.zimbra.cs.mailbox.IndexHelper.indexDeferredIte msInternal(IndexHelper.java:367)
at com.zimbra.cs.mailbox.IndexHelper.indexDeferredIte ms(IndexHelper.java:284)
at com.zimbra.cs.mailbox.IndexHelper.maybeIndexDeferr edItems(IndexHelper.java:240)
at com.zimbra.cs.mailbox.Mailbox.addMessage(Mailbox.j ava:4310)
at com.zimbra.cs.mailbox.Mailbox.addMessage(Mailbox.j ava:4303)
at com.zimbra.cs.mailbox.Mailbox.addMessage(Mailbox.j ava:4251)
at com.zimbra.cs.filter.IncomingMessageHandler.addMes sage(IncomingMessageHandler.java:113)
at com.zimbra.cs.filter.IncomingMessageHandler.implic itKeep(IncomingMessageHandler.java:106)
at com.zimbra.cs.filter.ZimbraMailAdapter.doDefaultFi ling(ZimbraMailAdapter.java:281)
at com.zimbra.cs.filter.ZimbraMailAdapter.executeActi ons(ZimbraMailAdapter.java:185)
at org.apache.jsieve.SieveFactory.evaluate(SieveFacto ry.java:164)
at com.zimbra.cs.filter.RuleManager.applyRulesToIncom ingMessage(RuleManager.java:348)
[...]

After several deliver tries with the same error, the Zimbra system crashes and we have to restart it in order to have the service up again.

I can access this problematic mailbox form any IMAP client even from the Zimbra webmail. I can even copy it to another mailbox using the imapsync tool, but the problem persists in the new mailbox.

It is very surprising to me that a problem in one mailbox may originate the failure of all the Zimbra system.

How can I solve this problem?
Any help is welcome.

Thank you very much for your help.
Best regards.

[phoenix]

So, how much RAM is on this system?

[uxbod]

As Phoenix has asked how much physical RAM plus the following

free
su - zimbra
zmlocalconfig | grep -i mailboxd_java

[soltecsis]

Our server has 8GB of RAM:
[...]
# free
total used free shared buffers cached
Mem: 8190040 5965752 2224288 0 359652 2759216
-/+ buffers/cache: 2846884 5343156
Swap: 3903752 0 3903752
[...]

[...]
$ zmlocalconfig | grep -i mailboxd_java
mailboxd_java_heap_memory_percent = 40
mailboxd_java_heap_new_size_percent = 25
mailboxd_java_options = -server -Djava.awt.headless=true -Dsun.net.inetaddr.ttl=60 -XX:+UseConcMarkSweepGC -XX:NewRatio=2 -XX:PermSize=128m -XX:MaxPermSize=128m -XX:SoftRefLRUPolicyMSPerMB=1 -verbose:gc -XX:+PrintGCDetails -XX:+PrintGCTimeStamps -XX:+PrintGCApplicationStoppedTime
[...]

[xtrenge]

same here, i got:

$ free
total used free shared buffers cached
Mem: 4058256 4029124 29132 0 30616 221140
-/+ buffers/cache: 3777368 280888
Swap: 3035128 111736 2923392


$ zmlocalconfig | grep -i mailboxd_java
mailboxd_java_heap_new_size_percent = 25
mailboxd_java_heap_size = 998
mailboxd_java_options = -server -Djava.awt.headless=true -Dsun.net.inetaddr.ttl=${networkaddress_cache_ttl} -XX:+UseConcMarkSweepGC -XX:PermSize=128m -XX:MaxPermSize=128m -XX:SoftRefLRUPolicyMSPerMB=1 -verbose:gc -XX:+PrintGCDetails -XX:+PrintGCTimeStamps -XX:+PrintGCApplicationStoppedTime -XX:-OmitStackTraceInFastThrow


we think that is problem of spam but... still with downs of zimbra, if someone know about this, please help i run out of ideas..

Thanks and happy new year