Thread Panic boLength(bo)

[Michael Hartzog]

We are on a SunFire V245 cluster running Solaris 5.9 and Cloverleaf 5.5Rev1. Our site has 14 processes

Topic

[Russ Ross]

This sounded familiar to me and I see this has been discussed before on clovertech at this URL:

<a href="http://clovertech.infor.com/viewtopic.php?t=1530&&#8221; class=”bbcode_url”>http://clovertech.infor.com/viewtopic.php?t=1530&amp;

The comment that caught my attention at that post was that the SMAT file might be too large.

If that is correct you might want to check all your SMAT files to see if any are very big.

If it is a SMAT problem you might also want to check all your interfaces to be certain that you haven’t configured 2 interfaces to write to the same SMAT file which is another common problem that can cause SMAT cycling a challenge.

Russ Ross
RussRoss318@gmail.com

[Kevin Scantlan]

I ran into the same message in a process log after it had aborted.  Based upon what you said to look for, I checked the process and found that the SMAT file for the ACK messages on a thread in that process was 2GB for each the .idx and .msg files.  I deleted them and was able to bring the process back up without any problems.

Thanks!

[Author]

Replies