starting swagentd

[martin wells_1]

Hi all,

I'm trying to register a depot locally on a HPUX 11.0 system but getting the error



the swagentd is not running and I try to start it with /sbin/init.d/swagentd start and nothing seems to happen.
ERROR: RPC exception: "Communications failure (dce / rpc)" 12/21/04
12:29:09 GMT
ERROR: A Remote Procedure Call to a daemon has failed. Could not
start a management session for
"ukbsfspd:/var/adm/crash/Sendmail-811_B.11.00.01.005_HP-UX_B.11.00_32+64.depot".
Make sure the host is accessible from the network, and that
its daemon, swagentd, is running.



There does not seem to be any problems with resolving the ip and host name. the SD_UX patch is a bit old but is the same level as the other 10 identical servers I've already registered the depot on?

thanks

[Robert-Jan Goossens]

Hi Martin,

try a nslookup.

# nslookup ukbsfspd
# nslookup ip_ukbsfspd

if the ip or hostname is not correct, you will have to change it in the /etc/hosts file.

# swagentd -r
to restart the agent.

Hope this helps,
Robert-Jan

[Steven E. Protter]

Maek sure you have a valid /etc/nsswitch.conf file on the system.

Then restart the daemon again.

There are several candiates in /etc but 11.00 and 11i v1 systems install without an actual nsswitch.conf file, without which swagentd can't function.

SEP

[martin wells_1]

I can resolve the name with nslookup, should i be able to resolve "ip_hostname"??

I've also got a /etc/nsswitch.conf file that seems to be the same as the other servers

[Jan Sladky]

Hi Martin,

I had same problem after new HP-UX installation.
The changes in /etc/hosts (ip and hostname)
and setting of /etc/nsswitch.conf helped

don't forget restart swagent -r


rgds Jan

[Eric Antunes]

Hi,

Check this thread because it seems to be similar to your issue: http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=767682

Regards

Eric Antunes

[martin wells_1]

Thanks Guys,

Still can't seem to get things working.

Should the command

nslookup ip_ukbsfspd

come back with something?

[Robert-Jan Goossens]

Nope :-)

I don't know the ip address of your server.

# grep the_name_of_server /etc/hosts

# nslookup above ip_address

should match with

# nslookup the_name_of_server

Hope this helps,
Robert-Jan

[martin wells_1]

Sorry Robert-Jan it's me getting confused

yes they do match?

[Robert-Jan Goossens]

example
hostname=pluto
ip=125.2.4.12

if there is an error in the naming or addressing somewhere you can track it with nslookup.

nslookup pluto
nslookup 125.2.4.12

should match, if they don't check the /etc/nsswitch.conf

hosts files dns

nslookup will look first in the files (etc/hosts) then dns

check the /etc/hosts

first name will be the local server.

125.2.4.12 pluto pluto.mars.com

Robert-Jan

[Mohanasundaram_1]

HiMartin,

If you had created the nsswitch.conf file after the system booted, then try moving that file to another name and try swlist.

Check if the filesets are in configured state. Otherwise get them to configured state first.

If you modified your hostname after this boot, then too such problem can come. check the outputs of uname and hostname and confirm that they are matching.

was the IP address modified ? If so, just do a restart.

ukbsfspd:/var/adm/crash/Sendmail-811_B.11.00.01.005_HP-UX_B.11.00_32+64.depot

this is a strange error. have you put any depot in the crash directory? though it may not be related to the problem, move this depot to another location. this is a very weird message.

Is any of your filesystem 100% full?

Check these things and revert.

With regards,
Mohan.

[martin wells_1]

Everything looks ok from the resolution side.

The error I posted at the start of the thread is the one that comes from the swreg command and not when I try and start the swagentd.

ps -ef for swagentd shows its not running

/sbin/init.d/swagentd start just comes straight back exit 0 but does nothing?

The nsswitch file was not created by me and i have not altered the hostname. Also all patches seem to be applied,

thanks

[Robert-Jan Goossens]

Ok,

# ls -la /etc | grep nsswitch

nsswitch.compat
nsswitch.files
nsswitch.hp_defaults
nsswitch.nis
nsswitch.nis+

will get a simular output as above.

# cp /etc/nsswitch.files /etc/nsswitch.conf

# swagentd -r
to restart to swagentd.

Robert-Jan

[martin wells_1]

Robert-Jan

The ls -al of /etc does show the same as above. However the two files are already the same?

Thanks

[Robert-Jan Goossens]

could you post the error from the

# tail -20 /var/adm/sw/swagentd.log

Roobert-Jan

[martin wells_1]

The last entry in the log is from 23/10/04??

[martin wells_1]

The entry anyway?

tail -20 /var/adm/sw/swagentd.log
* Stopping swagentd program, process ID 1092.

======= 10/23/04 08:05:15 BST END swagentd (pid = 15208).


======= 10/23/04 08:05:15 BST END swagentd (pid = 1092).


======= 10/23/04 08:15:17 BST BEGIN swagentd (pid = 369).

* Started remove agent on "/" for root@ukbsfspd, pid=377,
10/23/04 08:15:19 BST
* Agent pid=377 completed. 10/23/04 08:15:21 BST

======= 10/23/04 08:16:19 BST BEGIN swagentd (pid = 1097).

* Restarting swagentd program, process ID 369.

======= 10/23/04 08:16:19 BST END swagentd (pid = 369).

[Robert-Jan Goossens]

Just a thought

# ls -la /usr/sbin/swagentd

# file /usr/sbin/swagentd

what is the time/date stamp on the file and what kind of file is it ?

Robert-Jan

[martin wells_1]

ukbsfspd:/var/adm/sw# file /usr/sbin/swagentd
/usr/sbin/swagentd: s800 shared executable dynamically linked

ukbsfspd:/var/adm/sw# ll /usr/sbin/swagentd
-r-xr-xr-x 1 bin bin 491520 Apr 29 1999 /usr/sbin/swagentd

??

Thanks

[Robert-Jan Goossens · Accepted solution]

looks good.

how abou if you execute the binary with no options.

# /usr/sbin/swagentd

Robert-Jan

[martin wells_1]

It started it??

It must be a problem with the script in init.d in that it never runs the start?


Thanks for you help Robert-Jan!