Hi all,

I start this vote trying to solve my veto. I didn't do that before because I didn't understand someone wanted to make a 2.3.1 release in the past weeks.

The issue is:
OOM caused by unbounded cache in InetAddress (was James leaks memory slowly)
https://issues.apache.org/jira/browse/JAMES-592

Noel made an ad-hoc fix for the 2.3 branch that you can see here:
http://www.mail-archive.com/[email protected]/msg11380.html

I suggest to read again the whole thread before voting.

In the mean time I fixed the following DNSServer related bugs in Trunk:

- DNSServer does not properly cache MX lookups for CNAMEd domains
https://issues.apache.org/jira/browse/JAMES-672

I consider this not a critical issue, we can avoid backporting.

- DNSJava cache not properly handled when using multiple dnsjava components
https://issues.apache.org/jira/browse/JAMES-674

This is minor, doesn't deserve backporting.

- fetchmail cannot complete processing, because of problem with dnsjava and firewall/dns
https://issues.apache.org/jira/browse/JAMES-657

I didn't understand why my fixes fixed this problem, but if this is the case we should consider if this deserve more attention for 2.3.1 or if we want to ignore it.

- Make sure our container use an expiration for cached dns data
https://issues.apache.org/jira/browse/JAMES-679

This one is a workaround that would avoid JAMES-592 to happen, with no changes to the code. Noel suggested the use of "java.security.Security.setProperty("networkaddress.cache.ttl" , "0"); " instead of the -Dsun.net.inetaddr.ttl=0, but I expected some reply to the last comment there.
Please reply also there if you have preferences about this.

With this vote I'm trying to fix the v2.3 branch backport for 2.3.1 release (so this is not necessarily related to next-minor). This is why I remove the "full backport from trunk" from the options, because as Noel pointed out, we would break assembly.xml compatibility and we cannot do this in a point release.

And here are the options to be voted:

1) Backport the custom static access to DNSServer.
http://svn.apache.org/viewvc?view=rev&revision=470929

2) Backport sun jvm specific fix to JAMES-679:
http://svn.apache.org/viewvc?view=rev&revision=473041

3) Define where to add the "java.security.Security.setProperty("networkaddress.cache.ttl" , "0");" apply it to trunk and backport it (if you vote this please tell your preferences about the place to add this code).

Please note that every single of this 3 solutions will remove this OOM so I think we should simply choose one solution to be backported.
If you agree on this we'll take the solution with the best average vote.

It also worth saying that while 1 is a fix for the problem, 2 and 3 are workarounds to avoid the OOM, but less optimized as solutions.

Stefano


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to