[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference turris::digital_unix

Title:DIGITAL UNIX(FORMERLY KNOWN AS DEC OSF/1)
Notice:Welcome to the Digital UNIX Conference
Moderator:SMURF::DENHAM
Created:Thu Mar 16 1995
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:10068
Total number of notes:35879

8658.0. "DTAdvFS Utility Hanging With DECthreads Bugcheck" by NETRIX::"[email protected]" (Alexander Williams) Thu Jan 30 1997 17:15

I have a customer that is experiencing hangs when accessing dtadvfs.  In the 
/var/opt/advfsd/logs/advfsd file can be seen the following lines:

===
send_comm_v2_trap: Sendto failed, unable to send trap to manager.
at line 923 in file
/advfs_gui/sandboxes/megabld/pta.sandbox/snmp/_alpha-osf1.b/../snmpd/unix
   /trap.c
DECthreads bugcheck version 3.13-435, terminating execution
vpUpcallThreadUnblocked: (os/kern) invalid argument (4) nxm_reshed (22,0)
===

It /looks/ as if dtadvfs is having problems calling the advfsd via thread
call; 
has anyone seen this before or does Engineering need to get a nice, concise 
data-package?  :)

Alexander
[Posted by WWW Notes gateway]
T.RTitleUserPersonal
Name		DateLines
8658.1DECWET::MARTINThu Jan 30 1997 19:5715
Is the advfsd process still running?  If advfsd died a heinous death, dtadvfs
will appear to hang.  (It'll eventually time-out, I believe and give you a fatal
error to exit, but it takes somewhere around 30 minutes to time out.)

Looks to me like DECthreads bit the SNMP thread in the butt, or the SNMP thread
choked making it appear to be a DECthreads problem.  Dunno which.

Is this repeatable?  If so, we need all the info we can get on it.  We've seen
many, many cases similar to this (there's probably a release note on this
problem), but have never been able to create a reproducible scenario so that we
could figure out what was going wrong.

Please feel free to file a CLD/IPMT/QAR/whatever you feel is appropriate.

--Ken
8658.2AdvFS Biting SNMP, News at 11NETRIX::"[email protected]"Alexander WilliamsFri Jan 31 1997 14:5610
As far as I can tell, its not only repeatable, its repeatable at will.  All
it takes is for this fine fellow to start up dtadvfs on his 8200 and forward
the DISPLAY to another box; every time it dies a heinous death.

I'll get him the latest Unicensus/check_it script and then IPMT it.

[bounce bounce]  I get to IPMT, yippie!


[Posted by WWW Notes gateway]
8658.3Do NOT IPMT this case...see QAR 51214 on GORGE, OSF_QAR databasePTHRED::PORTANTEPeter Portante, DTN 381-2261, (603)881-2261, MS ZKO2-3/Q18Fri Jan 31 1997 15:4013
Folks,

This is a known problem.  I have developed the fix for this bug.  I am testing
the patch now.  It will not be done today, so it will not make the February
patch kit.  I don't know when the next one will be.  The Digital UNIX Support
folks should be able to tell you.

See notes 1473.* in the DECthreads notes conference, CLT::CMA.

Thanks,

-Peter Portante
 DECthreads Developer