Ifxclone issue
User reported that after building a Replicated Secondary Server (RSS) using ifxclone utility on IDS 11.70.FC5X7, all onmode commands hang on the Primary instance, requiring manual kill of the oninit process. IBM support suggested this was likely a defect and recommended opening a PMR with diagnostic output (onstat commands) to investigate further. No resolution was provided in the thread.
Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.
Hi, A big thanks to Mr Madison Pruet for developing IFXCLONE utility. I have been using this for over a year now and it works fine on most occassions. Only issue I see is after the RSS is built, the Primary fails to respond to all ONMODE commands. My IDS version is 11.70.FC5X7 and OS is uname -a SunOS 5.10 Generic_147441-07 i86pc i386 i86pc Is this a known issue? regards, Nitin Mathur
any onmode commands? you can't roll logs forward or shut it down or put it in
standard mode? I've been using ifxclone about an equal amount of time and it's
not done anything strange like that here.
Yes, every onmode command hangs. I can't shutdown the Primary instance also, I
have kill the master oninit process.
Is it logging anything at all on the primary or secondary? is the primary still available for reads and writes or does it hang?
original post:
Yes, every onmode command hangs. I can't shutdown the Primary instance also, I
have kill the master oninit process.
response:
If all onmode commands are hanging and you have to bounce the server, it
sounds like it could be a defect (it's hard to say without seeing any onstat
output showing the various states of various threads). I'd recommend opening a
pmr for the issue, and if you can reproduce the problem, I'd reproduce it, and
at if possible, grab an onstat -a, onstat -g all, onstat -g stk all, and if
possible a shared memory dump (onstat -o) from the primary (or which ever
machine the onmode commands were hanging on), and submit the onstat output
(well all the onstat output but the onstat -o output, which may not be needed
but to be safe it's best to go ahead and collect it if it is required) when
the pmr was opened.
Jacques Renaut
IBM Informix Advanced Support
APD Team
Thanks Jacques.
I have seen this defect many times and currently I have the issue on my
production database. I haven't killed the master daemon as yet as I need to
wait for the weekend. However, I dont see anything else failing and no
complaints from the users/apps.
I will gather all onstat outputs as advised and share with IBM on PMR.
regards,
Nitin