axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From bugzi...@apache.org
Subject DO NOT REPLY [Bug 11532] - Out of Memory Error - tcpmon
Date Wed, 04 Dec 2002 20:56:59 GMT
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG 
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://nagoya.apache.org/bugzilla/show_bug.cgi?id=11532>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND 
INSERTED IN THE BUG DATABASE.

http://nagoya.apache.org/bugzilla/show_bug.cgi?id=11532

Out of Memory Error - tcpmon





------- Additional Comments From rkimbrell@northropgrumman.com  2002-12-04 20:56 -------
Glen,

Nothing new.  Our purpose in the large-packet benchmark was to prove that SOAP
would be a viable part of the enterprise we're building (a transaction-based
weather data system).  However, since SOAP attachments work just fine, we've no
particular need to send large packets.  - And by using SOAP attachments, we
proved our point and will be using it.

I believe this "out of memory" problem isn't peculiar to Axis - it seems to have
occurred on our Sun SunOne box as well - although who knows the geneisis of the
SunOne SOAP code.  It might well be Axis, too.

In any case, this problem appears to be a design issue, though it would be nice
for the failure to be a bit more controlable.  Here are my thoughts.  At some
point the JRE is going to run out of memory.  And then, well, there's no more
memory for large packets.  The benchmark shows what appears to be a memory leak
(but that can't be it, of course) because even with small packets, we eventually
run out of memory.  I think what's happening is a problem with garbage
collection.  We're asking for memory faster than the garbage collector can
deliver it.  (It's a separate process, right?)  I don't think we want to be
messin' with the garbage collector since there are lots of ways to handle large
memory problems, and just requesting a large block of memory or several large
blocks is a sorta brute-force approach - saner minds should devise better
solutions.  SOAP attachmens is an example.

In summary, I think this is a problem that could use some attention just to
verify that it's what I suppose it is and, perhaps, devise a graceful response
to "out of memory" errors.  Also, if the problem is with the garbage collector,
then a bit of documentation and some alternative approaches to large memory
allocation might be the ticket.

Thanks much for all your help.  You guys are greatQ!

Regards,
Roy

Mime
View raw message