hciconnstatus errors

[Anonymous]

Happy New Year everyone!

Our server crashed last night – and it took me several hours to get it back up and get our interfaces functioning again.

Now, this morning, we find that on at least one of our sites, hciconnstatus doesn’t work.  We get the error:

bgerror failed to handle background error.

   Original error: unmatched open brace in list

   Error in bgerror: invalid command name “bgerror”

Error: can’t wait for variable “someBogusVariable”:  would wait forever

   while executing

“vwait someBogusVariable”

   (procedure “hciEventLoop” line 3)

   invoked from within

“hciEventLoop                   “

   (procedure “main” line 52)

   invoked from within

“main $argc $argv”

   (file “/qdx/qdx5.3/integrator/bin/hciconnstatus” line 149)

I have cleared the stats using hcimsiutil -R, and have even tried initializing the databases.  Any input on our error?

Topic

[Anonymous]

FYI – another reboot of the server fixed the problem.  Not sure what was hosed up, but the extra reboot got it.

[Ben Hochstedler]

I get the same error occassionally when running hciprocstatus (run once a minute to watchdog processes):

Quote:

bgerror failed to handle background error.

  Original error: broken pipe

  Error in bgerror: invalid command name “bgerror”

Although I don’t get the other error messages that hciconnstatus produced.  I too have found that rebooting is required to clear the condition.

Have you seen the problem happen again?

What OS did it happen on?

Any further clues what caused it?

I’m running:

Windows 2000 Server SP4

QDX 5.2

btw, I submitted another post before I found this thread.

[Scott Folley]

I am guessing somewhat but I believe what you have is a cmd_port file that still exists and hitting a port that is listening but is not an actual cloverleaf port.  A “proper” shutdown during the reboot cycle and the elimination of a process actually listening on that port is probably what is cleaning things up.  It is likely that you would do fine by doing a manual cleanup of all cmd_port, pid, and wpid files associated with the site.  Just my 2 cents, hope it helps.

[Ben Hochstedler]

Thanks for the reply.  I’ll have to look into that as the possible root cause.

I’ve identified two workarounds for the problem:

[Author]

Replies