Re: Forced Node Shutdown (error 2303)

[Stewart Smith]

On Mon, 2006-06-19 at 17:00 +0200, Dirk Dunger wrote:
> I am currently trying to write a transporter for NDB but I'm having some
> problems getting it done right. The nodes get connected via that
> transporter and seem to start as expected. But when I run testReadperf
> one node gets shut down after a while and I'm not quite sure what my
> transporters are doing wrong.
> I put the logs and tracefile on my webspace [1] hoping that someone with 
> more experience in interpreting the information in those files could 
> give me
> a hint.

basic way to read traces:

left to right, top to bottom:
left is block name (ndb/src/kernel/blocks/)
numbers are line numbers (jamEntry(), jam() - and various asserts)
where there are high line numbers (10000, 30000) there are probably more
than 1 source file for the block. source file (or header) will have a
#define ljam() or #define jam() with a 10000+__LINE__ or something.

below this is the recent signals.
r.bn Receiver Block Number
s.pn Sender " "
r/s.proc process ID (node)
gsn - global signal number
then the content of the signal.

so here you can see that you died in NDBCNTR, at line 195. the last
signal sent was SYSTEM_ERROR from DBDIH on node 2 to NDBCNTR on node 2.

You probably want to look into it from there.

my guess would be that there's a bug in your transporter corrupting
signals.

what transporter are you writing?

maybe try running under valgrind.
-- 
Stewart Smith, Software Engineer
MySQL AB, www.mysql.com
Office: [phone] Ext: 6616
VoIP: 6616@stripped
Mobile: [phone]

Jumpstart your cluster:
http://www.mysql.com/consulting/packaged/cluster.html

Attachment: [application/pgp-signature] This is a digitally signed message part signature.asc

Attachment: [application/pgp-signature]