swagentd errors

[Robin King]

I have been trying to install some software on one of our servers and have encountered a problem with the swagentd.

This problem has been logged to the Resource Centre, but was hoping someone here may be able to shed some light, as they're not having much luck.

The problem is that I am unable to start the swagentd. The command line error is:

hostname:/#/usr/sbin/swagentd -r
ERROR: swagentd exiting due to initialization error in backgroun
process. Check the logfile "/var/adm/sw/swagentd.log" fo
more information.

The entry in the swagentd.log shows the following:

======= 04/30/01 11:13:37 BST BEGIN swagentd (pid = 2746).

WARNING: Failure registering as Remote Procedure Call server using
protocol sequence "ncacn_ip_tcp" with endpoint "2121". status
0xfffffffe (??? / )
ERROR: Terminating prematurely - cannot inquire RPC bindings. status
0xfffffffe (??? / )

======= 04/30/01 11:13:37 BST END swagentd (pid = 2746, startup
error).

Has anyone experienced a problem similar to this that may be able to advise. As it is now becoming quite urgent that I get this problem sorted.

Cheers!!

[Vincenzo Restuccia]

Try:1) ps -ef | grep sw
That will pull up swagentd (usually swapper, but has nothing to do with this here and should be left alone)
2) kill -9 PID for swagentd
3) check for other sd ux processes and if so check if they can be killed
4) /usr/sbin/swagentd ... to restart the swagentd

[Robin King]

My problem isn't restarting the daemon as such it's actually starting it.
There is no swagentd daemon running at present.

hostname:/#ps -ef | grep sw
root 0 0 0 Aug 21 ? 22:09 swapper
root 4307 29249 1 11:41:02 pts/tf 0:00 grep sw

Robin

[Thierry Poels_1]

hi,
has the hostname been changed?
did you try a reboot?
is the hostname also in /etc/hosts?
regards,
Thierry.

[Vincenzo Restuccia]

See http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0xda9f0cb17a32d5118fef0090279cd0f9,00.html

[Vincenzo Restuccia]

Hi again,see also http://europe-support2.external.hp.com/cki/bin/doc.pl/sid=4fab6b8b1b83b1aab9/screen=ckiDisplayDocument?docId=200000024631715

[Robin King]

The hostname hasn't changed sinced the original build 2 years ago, and yes the hostname is in /etc/hosts.

The box hasn't been rebooted, as it isn't very easy to get downtime, the server has been up for 252 days and they like it that way ;o)
I could get permission for a reboot window, but wouldn't be until the middle of the night.

This was something HP asked. Just to let you know other things I have tried.....

(1) Apparently a similar problem to this can be caused by lan card failures, but the system only has one card and this is working fine.

(2) An "swlist -l fileset -a state | grep -e install -e trans -e corrupt" returns some filesets that are in a transient state. HP recommended configuring these (swconfig \ *), but this won't work, it again complains that the swagentd isn't running.

(3) Removed the /var/adm/sw/products/swlock file.

None of these have made any difference.

Cheers

[Rita C Workman]

Here's a thread that mentions that ncan_ip_tcp:
http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0xda9f0cb17a32d5118fef0090279cd0f9,00.html

I hope it helps...and I have just one more thought. swinstall (like other apps) does some resolving when it starts up...so if it can't resolve it has a problem. Now you mentioned that this system has not been changed in 2 years. BUT...is it possible that somehow a duplicate IP is floating out there somewhere. If your getting this error try doing a nslookup on both name and IP...see if they agree everytime. I've seen some pretty flakey things happen because somehow an IP got dup'd.

Like I said...just a thought,
/rcw