lsof on HP 11.11/11.23 PA

[derek b smith_1]

Hi All,

I looked through all the related threads here and I am having the exact problem, lsof hangs on pstat using tusc.

Will anyone send me a working lsof depot for 11.11 and 11.23 or show me where to attain one?
derek.smih@cbc-companies.com
Is there a patch, a hack or an older version I need?

v4.80 works on my 11.23PA servers, but I cannot find the working depot at my workplace and not all servers have it. Disgruntled former admin... Anyway, v4.83 on 11.23PA from Merlin, and the HP internet package does not work, it hangs as well.

thank you!

[Tingli]

Try,

http://hpux.connect.org.uk/

[derek b smith_1]

Tingli from CAH years back?
I did try that source...same issues. :(

thanks anyway.

[Tingli]

Sorry for not much help.

I was in CAH few years ago.

[Dennis Handly · Accepted solution]

>I am having the exact problem, lsof hangs on pstat using tusc.

Hangs or spends a long time as in this thread?
fuser uses too much CPU

http://h30499.www3.hp.com/t5/System-Administration/fuser-uses-too-much-CPU/td-p/4606069

[derek b smith_1]

I believe it hangs because it does not come back. It hangs on:

pstat(PSTAT_SOCKET, 0x800003ffefff1118, 648, 0, 0) ......................................... [sleeping]


The utild daemon is not running nor is it installed on my 11.23PA host.

[Dennis Handly]

>It hangs on: pstat(PSTAT_SOCKET, 0x800003ffefff1118, 648, 0, 0)

Ah, this thread has a similar hang:

http://h30499.www3.hp.com/t5/General/lsof-4-81-issue/td-p/5160782

Are you running the same application?

[derek b smith_1]

No not same app. Actually this is on a admin server, but other prod servers are not using xppa or java. Most of our apps are in C or C++. The version of lsof on the prod server is 4.77, but same it hangs.

Please explain how to get the offending PID again, do I need to start lsof in a seperate session then run?

Try finding which PID has the socket? Loop through all of the PIDs with "lsof +aL1 /var -p PID1":
for P in $(UNIX95=EXTENDED_PS ps -e -opid=); do
lsof -p $P +aL1 /var
done

Here is my gdb data:

# gdb /usr/local/bin/lsof 14684
HP gdb 5.4.0 for PA-RISC 1.1 or 2.0 (narrow), HP-UX 11.00
and target hppa1.1-hp-hpux11.00.
Copyright 1986 - 2001 Free Software Foundation, Inc.
Hewlett-Packard Wildebeest 5.4.0 (based on GDB) is covered by the
GNU General Public License. Type "show copying" to see the conditions to
change it and/or distribute copies. Type "show warranty" for warranty/support.
..
warning: Load module /usr/local/bin/lsof has been stripped

(no debugging symbols found)...
Attaching to program: /usr/local/bin/lsof, process 14684
(no debugging symbols found)...(no debugging symbols found)...(no debugging symbols found)...
(no debugging symbols found)...(no debugging symbols found)...
warning: reading `r3' register: No data
0xc02167d0 in _read_sys+0x10 ()
from /usr/lib/libc.2
(gdb) where
#0 0xc02167d0 in _read_sys+0x10 () from /usr/lib/libc.2
#1 0xc0221b1c in read+0x6c () from /usr/lib/libc.2
#2 0xf4a8 in + 0x210 ()
#3 0xf9f4 in + 0x1bc ()
#4 0x1739c in + 0xbc ()
#5 0x39f8 in + 0x370 ()
#6 0x52e4 in + 0x41c ()
#7 0xd3f8 in + 0x1b38 ()

[Dennis Handly]

>No not same app.

Your stack trace shows it is hanging in a read, not pstat.

What lsof option were you using?

>Please explain how to get the offending PID again, do I need to start lsof in a separate session then run?

Take your lsof command and add that -p $P to it and put it in that loop.
You don't need to be a separate session.

My later reply suggested you put this in the loop so you know which PID hangs:
for P in $(UNIX95=EXTENDED_PS ps -e -opid=); do
    UNIX95=EXTENDED_PS ps -x -p $P -opid= -oargs=
    lsof -p $P +aL1 /var # adjust as needed
done

>Here is my gdb data:
#0 0xc02167d0 in _read_sys+0x10 /usr/lib/libc.2

This is hanging in a read, not pstat. It may be interesting to use tusc on lsof to see what it was doing to get to that read?

Was it hung before you attached with gdb?

[derek b smith_1]

Any option or no option for lsof, but regardless I found the culprit PID. It was a BMC scheduling process on all but one host.
./ctmagt/ctm/exe/p_ctmat

Stopping and restarting this BMC solved the problem, but this reverse library PID I will take extra caution.

On the loner host its a library file called /usr/lib/reverse

# what /usr/lib/reverse
/usr/lib/reverse:
$Revision: vw: cheriv_JAGad35417 selectors: CUPI80_BL2000_1025 'cup_cheriv_jagad35417'
Fri Oct 27 00:29:24 PDT 2000 $

thank you much!!!