Bugs item #1007959, was opened at 2004-08-12 15:39
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=376685&aid=1007959&group_id=22866

Category: Clustering
Group: v3.2
Status: Open
Resolution: None
Priority: 5
Submitted By: Peter Backx (pbackx)
Assigned to: Nobody/Anonymous (nobody)
Summary: HA-JNDI extremely unreliable

Initial Comment:
os: linux 2.6
jdk: Sun hotspot 1.4.2_02

I have deployed 2 SLSBs on 2 different machines that
are clustered (I tried both TCP and UDP) HA-JNDI is
enabled but no farming.

Both looking for a bean from the client and looking for
a bean from another is extremely unreliable. It works
in about 10% of the cases. I have been unable to make
this bug 100% reproducable ... it appears to be
entirely random.

If I let one bean call another I get the following in
my server logs which shows that the server is actually
responding:

2004-08-12 15:31:08,759 DEBUG
[org.jboss.ha.framework.server.HAPartitionImpl]
dests=[asklepios30:7800 (additional data: 16 bytes),
10.10.12.1:7800 (additional data: 15 bytes)],
method_call=HAJNDI.lookupLocally(Another), mode=2,
timeout=60000
2004-08-12 15:31:08,759 DEBUG
[org.jboss.ha.framework.server.HAPartitionImpl]
real_dests=[asklepios30:7800 (additional data: 16
bytes), 10.10.12.1:7800 (additional data: 15 bytes)]
2004-08-12 15:31:09,027 DEBUG
[org.jboss.ha.framework.server.HAPartitionImpl]
responses: [sender=asklepios30:7800 (additional data:
16 bytes), retval=null, received=true, suspected=false]
[sender=10.10.12.1:7800 (additional data: 15 bytes),
[EMAIL PROTECTED],
received=true, suspected=false]

(I have 3 servers: 10.10.11.1 is doing the lookup and
asklepios30 and 10.10.12.1 are replying ... don't mind
the fairly complicated (for a cluster) network setup,
the problem is the same if they are all on the same
subnet and UDP multicasting)

However by the time this value gets returned by
Context.lookup it is null (most of the time).

This bug was not in 3.2.2 but because I cannot assign
the correct NodeAddress to partition members I have to
use 3.2.5

I have also attached my cluster configuration file
(@BIND_ADDR@ gets replaced by the correct address in an
ANT script)

I'm willing to debug the code myself because this
problem is really urgent to me, however I hope some one
can tell me through which calls the information flows
between the HAPartitionImpl and the lookup return. Any
pointers are appreciated.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=376685&aid=1007959&group_id=22866


-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
_______________________________________________
JBoss-Development mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-development

Reply via email to