Stm refuse to start

[Charles Castelain_1]

Hello,

I have installed hp-ux 11.11 dec 2004 on my C3600.
But no way to start stm on the system.

I am always receiving the following :

Unable to start the STM User Interface.
Failed to initialize the main window.

Possible Causes:

Internal Application error.

This is pretty uncorfortable of course.

I am also unable to start the diagnostic tools.
Is taht linked?

Regards,

Charles

[Robert-Jan Goossens_1]

Hi Charles,

Yes stm and diag are "linked".

Try to download and install the latest version.

http://www.software.hp.com/portal/swdepot/displayProductInfo.do?productNumber=B6191AAE

Regards,
Robert-Jan

[Steven E. Protter]

I had the same problem on L2000 boxes. It was solved by downloading the application and installing the later version.

swlist -l product | grep -i stm

maybe change stm to diag

SEP

[Charles Castelain_1]

Hello,

I did try to reinstall since I am already running the latest version.
No luck.
I try to launched manually diagmond.
Sounds nothing append... no crash no info no output...

Any idea?

For what I have understood xstm/cstm/mstm/stm will not work until the diags are running...

But I have no clue on how to make them working...

Regards,

Charles

[erics_1]

Charles,

What's the output of the following?

swlist -l fileset|grep Mgr

Also, have you tried swremove of diags and followed with a clean install? Sometimes, diag upgrades can leave mismatched filesets that will cause problems.

Regards,
Eric

[Charles Castelain_1]

Hello,

I did not try a remove/reinstall since this is a fresh install on a new drive.

Here is the output of swlist :

EMS-Core.EMS-MX A.04.00.02 Event Monitoring Services SCMgr Tools
# Sup-Tool-Mgr B.11.11.15.13 Support Tools Manager for HPUX systems
Sup-Tool-Mgr.RELEASE-NOTES B.11.11.15.13 HPUX Support Tools Manager Release Notes
Sup-Tool-Mgr.STM-CATALOGS B.11.11.15.13 HPUX Support Tools Manager Catalogs
Sup-Tool-Mgr.STM-MAN B.11.11.15.13 HPUX Support Tools Manager Manual Pages
Sup-Tool-Mgr.STM-SHLIBS B.11.11.15.13 HPUX Support Tools Manager Share Libraries
Sup-Tool-Mgr.STM-UI-RUN B.11.11.15.13 HPUX Support Tools Manager User Interface
Sup-Tool-Mgr.STM-UUT-RUN B.11.11.15.13 HPUX Support Tools Manager Unit Under Test Runtime

Regards,

Charles

[Denver Osborn]

what's the output of...

swlist -l fileset -a state Sup-Tool\*


If anything shows other than configured, try...

swconfig Sup-Tool\*

Then try starting up diags and running cstm/stm again.

If that still fails, I'd like to know what's wrong w/ diagmond not running... if you get this far w/ the problem still, I'd be curious to see the output from

sh -x /sbin/init.d/diagnostic start

Have you attempted to manually run
/usr/sbin/stm/uut/bin/sys/diagmond and get the same results?

hope this helps,
-denver

[Charles Castelain_1]

Hello,

Here is the output of swlist -l fileset -a state Sup-Tool\* :
# Sup-Tool-Mgr
Sup-Tool-Mgr.RELEASE-NOTES configured
Sup-Tool-Mgr.STM-CATALOGS configured
Sup-Tool-Mgr.STM-MAN configured
Sup-Tool-Mgr.STM-SHLIBS configured
Sup-Tool-Mgr.STM-UI-RUN configured
Sup-Tool-Mgr.STM-UUT-RUN configured

So nothing special...

For the sh -x /sbin/init.d/diagnostic start
I have joined the output.

When I try to launched diagmond manually nothing append... Really nothing no error no crash nor core not even a debug or a warning.

Charles

[Bill Hassell]

You are missing several patches that are related to hardware as well as the STM interfaces. As a minimu, add all the HWE (hardware enablement) patches from your 2005 SupportPlus CDROM. If you don't have that (no software support from HP), download the latest SupportPlus patch set from http://software.hp.com (you'll need both the Quality Pack as well as the hardware set.

[Charles Castelain_1]

Hello,

At the moment I can't access that site.
Receive gateway timeout.
But via itrc and patches the latest available is Decemeber 2004. I am alrady running that version.

Strange.

Regards,

Charles

[Ermin Borovac]

Any entries in /var/adm/syslog/syslog.log related to STM (diagmond etc)?

[Charles Castelain_1]

Hello,

Yes there are some :

May 20 16:37:45 hpwks01 diagmond[2642]: started
May 20 16:37:45 hpwks01 diagmond[2642]: Exit due to initialization failure
May 20 16:57:54 hpwks01 diagmond[5053]: started
May 20 16:57:54 hpwks01 diagmond[5053]: Exit due to initialization failure
May 20 16:58:19 hpwks01 diagmond[5091]: started
May 20 16:58:19 hpwks01 diagmond[5091]: Exit due to initialization failure
May 23 09:20:13 hpwks01 diagmond[11791]: started
May 23 09:20:13 hpwks01 diagmond[11791]: Exit due to initialization failure
May 23 09:20:51 hpwks01 diagmond[11808]: started
May 23 09:20:51 hpwks01 diagmond[11808]: Exit due to initialization failure

Regards,

Charles

[Bill Hassell]

Ahh, maske sure that registrar is enabled in /etc/services and /etc/inetd.conf. This is the STM/EMS communication port:

/etc/services:
registrar 1712/udp # resource monitoring service

/etc/inetd.conf:
registrar stream tcp nowait root /etc/opt/resmon/lbin/registrar /etc/opt/resmon/lbin/registrar

Once both entries are uncommented, run inetd -c and then restart diagnostic and ems:

/sbin/init.d/diagnostic stop
/sbin/init.d/ems stop
/sbin/init.d/diagnostic start
/sbin/init.d/ems start

[Charles Castelain_1]

Hello,

registrar was there but not properly configured since I have the IPV6 option.
But I am not using IPV6 yet.

# cat /etc/services | grep registrar
registrar 1712/tcp # resource monitoring service
registrar 1712/udp # resource monitoring service

# cat /etc/inetd.conf | grep registrar
#registrar stream tcp6 nowait root /etc/opt/resmon/lbin/registrar /etc/opt/resmon/lbin/registrar
registrar stream tcp nowait root /etc/opt/resmon/lbin/registrar /etc/opt/resmon/lbin/registrar

I have corrected it to use IPV4 not IPV6

so :

# netstat -an | grep 1712
tcp 0 0 *.1712 *.* LISTEN

So registrar is now working.
But diagmond still refuse to start.
Same error as previous.

Regards,

Charles

[Denver Osborn]

Have you also checked that port 1508 is available for diagmond and not being used by some other app? Make sure nothing else is already using port 1508 before starting up diagnostics. if you have lsof "lsof -i :1508" else "netstat -an"

If you have tusc installed, use tusc to start diagmond to see what's causing the problem.

tusc -fnvo /tmp/tusc.out /usr/sbin/stm/uut/bin/sys/diagmond


-denver

[Andrew Merritt_2]

Hi Charles,
Sounds like the OnlineDiags are failing to start up for some reason. I don't have an answer at the moment, but there's some data that might give some clues:

Did you have a previous version of OnlineDiags installed before? When did the problem start, was it with the installation or after that?

What is the output of:
ps -ef | grep -E 'stm|resmon|logd|psm'

If the problem happened around the upgrade time, was anything logged in /var/adm/sw/swagent.log?

Are there any error messages logged in api.log or client.log, or registrar.log in /etc/opt/resmon/log?

What happens if you run:
/sbin/init.d/diagnostic stop
followed by
/sbin/init.d/diagnostic start
?

Andrew

[Charles Castelain_1]

Hello,

Nothing is running on 1508.

But under /etc/opt/resmon/log I have a bunch of errors...

On the client it says in loop :

-------------------Start Event--------------------
Event 3 occurred at Tue May 24 17:33:53.678104 2005
Process ID: 3248 (/etc/opt/resmon/lbin/p_client) Log Level: Error
rm_client_connect: Cannot get IP address for hpwks01
-------------------End Event----------------------

-------------------Start Event--------------------
Event 2505 occurred at Tue May 24 17:33:53.680342 2005
Process ID: 3248 (/etc/opt/resmon/lbin/p_client) Log Level: Error
update_monitors: socket open failed
-------------------End Event----------------------

But a ping is working nice to that system...

Regards,

Charles

[Andrew Merritt_2]

Sounds like it's a networking configuration problem. What do you mean when you say you have IPV6 but are not using it?

Does 'nslookup' return the IP address for hpwks01?

Andrew

[Charles Castelain_1]

Hello,

I notice I forgot some answer :

# ps -ef | grep -E 'stm|resmon|logd|psm'
root 980 1 0 17:42:25 ? 0:00 /usr/sbin/syslogd -D
root 3237 1 0 17:44:01 ? 0:00 /etc/opt/resmon/lbin/p_client
root 2852 1 0 17:43:25 ? 0:00 /etc/opt/resmon/lbin/emsagent


This a fresh install and all the stm is marked as configured in swlist.

Latest error massage in api.log :
-------------------Start Event--------------------
User event occurred at Tue May 3 11:26:11.125017 2005
Process ID: 3862 (/usr/sbin/stm/uut/bin/tools/.../lpmc_em) Log Level: Error
The monitor cannot determine the port number of diaglogd to communicate for registration.

Possible Causes/Recommended Action:

It may be possible that diaglogd is not running on your system. It is also possible that all the monitors will experience the same problem till diaglogd is successfully launched. diaglogd failure messages will be logged only once for all the monitors.

Please shutdown and restart diaglogd using STM. On successful registration of the monitor, a message will be logged indicating the fact.
-------------------End Event----------------------

Latest error message in registrar.log :
-------------------Start Event--------------------
Event 2960 occurred at Tue May 3 11:08:57.000364 2005
Process ID: 3081 (/etc/opt/resmon/lbin/registrar) Log Level: Error
process_time_event: Expired awaiting-reply object, socket=25
-------------------End Event----------------------

-------------------Start Event--------------------
Event 2937 occurred at Tue May 3 11:08:57.001673 2005
Process ID: 3081 (/etc/opt/resmon/lbin/registrar) Log Level: Error
abort_awaiting_reply_obj: socket=25: Connection aborted
-------------------End Event----------------------

-------------------Start Event--------------------
Event 2961 occurred at Tue May 3 11:08:57.002756 2005
Process ID: 3081 (/etc/opt/resmon/lbin/registrar) Log Level: Error
process_time_event: Expired contact object for monitor /usr/sbin/stm/uut/bin/tools/monitor/fc60mon
-------------------End Event----------------------

You get the error of client log in the previous message.

Regards,

Charles

[Charles Castelain_1]

Hello,

The bundle with the extension for IPv6 is installed but I don't use currently an IPv6 address.

The ping and the nslookup is just fine.

# nslookup hpwks01
Using /etc/hosts on: hpwks01

looking up FILES
Name: hpwks01
Address: x.x.x.x
Aliases: hpwks01.mydomain.myext

In /etc/hosts I have :

x.x.x.x hpwks01 hpwks01.mydomain.myext

And in /etc/nsswitch.conf

hosts: files dns [NOTFOUND=return UNAVAIL=continue TRYAGAIN=return]

And in /etc/resolv.conf

domain mydomain.myext
nameserver x.x.x.x

Regards,

Charles

[Andrew Merritt_2]

Another thing to check is what version of EMS do you have installed? You should have A.04.00.02.

Andrew

[Andrew Merritt_2]

If the cause of the problem is not obvious, I'd recommend opening a support call with HP.

Andrew

[Charles Castelain_1]

Hello,

EMS is at the correct version.

I don't think I can open a call at hp since I have no contract for that system.

Regards,

Charles

[Andrew Merritt_2]

Ok, I think there's two things.

First, registrar uses the 'tcp6' version of the line, even on systems with IPv4.

Second, I think the problem is something in the configuration of the system that is stopping diagmond from initialising correctly.

Does the file /var/stm/logs/sys/activity_log exist, and is it updated each time 'diagmond' starts up (e.g. by running '/sbin/init.d/diagnostic start')? (This will indicate how far diagmond is getting in its initialisation.)

What is the status of IPCS on the system? Check with 'ipcs -mqs -ba' what the status is. One of the early things that diagmond does is to allocate ports and semaphores.

If there is an activity_log file, you could try running 'strings' on it to see if that shows anything useful, or if you have another system where STM is running successfully, you could try copying the file over and running the 'sal' command in cstm to see if diagmond is logging a failure there before it exits.

Andrew

[Charles Castelain_1]

Hello,

I would like to thanks all of you.
For your answers that help me to understand the process.
Understand that there were a "network" something error.
For your time and patience because this a very "special" case.

But I finally get it working and the solution is simply unbelievable...

In have configured everything to work with files and dns.
All the resolv system was working in files.
But since it is "a personal wks for my own pleasure and tests" I did not asked a entry in the central dns.
No reason to create an entry since I am the only user for that box.

I have verified with a server I have configured and that worked nicely...

The only difference was that server has a dns entry.

Without believing it would change a thing I have asked a dns entry...

Wait 10 minutes.
Logs starts to change under /etc/opt/resmon/log.

And believe it or not... now everything is working fine...

Regards,

Charles