I think I've found several issues with 2.1.0rc2
Environment:
RedHat Enterprise 4 Update 7
gcc version 3.4.6 20060404 (Red Hat 3.4.6-9)
Build Problems:
* GCC doesn't like the -R linking option removing
-R$ac_cv_path_mysql/lib/mysql from aclocal.m4 makes gcc happy.
The cleanup of junk files while building the apache2 filter (during a
make everything) causes cgi/logout to be rebuilt on 'make install' --
which fails because the cleanup also removes the libsnet libraries. IMHO
the following lines in filters/apache2/Makefile.in
rm -rf ${APXS2JUNKDIRS}
rm -f ${APXS2JUNKFILES
really belong in the clean target not the target for the DSO.
Replication Problems:
The other problem is more of a design issue with replication. We're just
starting to deploy cosign so our transaction rate is rather low. The
problem is if no REGISTER or LOGIN is received before the
cosignnettimeout expires the daemon side (target of the replication -
not source) of the pusher process quits. The results in a 'CHILD XXX
transient failure' message because the socket read returns EOF - no
other error is logged BTW. The next login/check request fails to
propagate to the replica, resulting in user errors (user doesn't appear
to be logged in even though they are) if a CHECK is done against the
replica that didn't get the LOGIN/REGISTER due to a timeout.
Should the daemon side of pusher have a different timeout? Or is
increasing the cosignnettimeout safe?
Thanks
Paul
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Cosign-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/cosign-discuss
