Processes are not coming up in Cloverleaf 6.1

[Varun Sinha]

I am trying to start the process and it’s not coming up. At GUI, it’s showing RED. Though, when I check from backend, it shows that the process is running.

Host server is running.

Lockmgr is running.

hcimonitord is running.

I tried starting from GUI, but the processes are not running.

From backend:        hcienginerun -p

I don’t know what to check further so that process should come up.

Regards,

 Varun

Topic

[Vanessa Rivera]

What we do at our site is turn all the processes off in that site and once everything is down we turn on the process that is giving us the issue.   Once that process comes back up we slowly start to turn on the remaining processes one by one.

Hope this helps.

[Varun Sinha]

I tried starting processes one by one, but none of them are coming up.

Also, We have installed 6.2 on the same box. But it was running fine for 6.1. We haven’t made any process up in 6.2.

[Vanessa Rivera]

We are on 6.1 and that seems to work for our sites.  We have not moved to 6.2 yet.

[Peter Heggie]

Possibly there is a conflict with 6.2 – is the host server that is running associated with 6.1 or 6.2? And you are running the 6.1 client? Also are you UNIX – some configurations use file /etc/environment to set the software version in the user context.

Peter Heggie

[Mary Kobis]

Has your license key expired? Are you over your licensed limit?

[Keith McLeod]

Is there a ‘pid’ file located in the process directory that won’t start?

If the process is down and the pid file exist, delete it, and cmd_port and wpid if they exist.  Then try to restart the process.

Not sure if you have already tried this…. Hope it helps.

[Mary Kobis]

Or try hciprocstatus on the command

[Peter Heggie]

So just to confirm – when you use a command line command – the process starts. The issue is that the GUI Network Manager does not show it as running and the GUI Network Manager does not respond to start commands.

Peter Heggie

[Jared Parish]

Are there a large volume of messages in the recovery database?  

I’ve had this happen as well.  The issue at hand was there was a large volume of messages in recovery db.  When each thread was starting, it would check every message to see if it was owned by that thread.  I discovered this by turning up the logging level.

- Jared Parish

[Keith McLeod]

Speaking of Recovery Database, I did run into a site having a large number of messages in the recovery DB for a destination that no longer existed.  The circumstance was created by the removal of a thread where the source thread was in another process.  They had bounced the process they removed the thread from but not other processes that were pointing to the thread.

I was performing what I thought was a simple change only to find out the process wouldn’t restart on account of a large volume of messages in the recovery DB.

I ran hcidbdump -r and learned of the messages.  hcidbdump -r -C will give you a total count of messages in the recovery DB.  Usually this number is 100 or less if no queued messages on any given thread.

Food for thought….

[Peter Heggie]

I encountered that one!

Peter Heggie

[Varun Sinha]

Thanks for the input.

Since, Cloverleaf 6.1 and 6.2 were running on the same box, somehow hci2 user process (for version 6.2) was trying to open the test site of Cloverleaf 6.1. We need to kill this process in order to make test sites up for Cloverleaf 6.1 as well as 6.2 on same Linux box.

          I have no idea as how this might have happened?

Regards,

 Varun

[Peter Heggie]

From a Linux command line, under the userid associated with Cloverleaf (hci ?), what is the result of a ‘showroot’ command? Is it 6.1 or 6.2 ?

When you do a command line command of ‘ps -ef | grep hci | grep java’, does it show a 6.1 or 6.2 process?

Peter Heggie

[Varun Sinha]

Hi Peter,

[Peter Heggie]

ok thank you. That one 6.2 monitor process is ok or not very bad, but it should not be there, and because I see it there, I am thinking there may be a problem somewhere else.

what is the output of showroot?

The attachment you included had only a few hci processes showing. Was there any other hci processes?

I did not see the host server. Can you issue this command: hciss

If it is not running, then just kill the 6.2 process kill -9.

Then make sure your environment is correctly pointing to 6.1

Peter Heggie

[Peter Heggie]

Also please check your Host Server environment

> hciss ?

It will show you the version of Host Server that is referenced in the current context. if it is 6.2, then there is an issue with your .profile or .profile.local.end files, or possibly the /etc/environment file (if your installation uses this file).

Peter Heggie

[Varun Sinha]

Hi Peter,

hciss was showing the correct 6.1.

Issue was resolved as you mentioned above.

We issued the command to kill the 6.2 process by kill -9 and then issued hciss -s h and it started the host server environment.

Currently, Cloverleaf 6.1 & 6.2 are running parallel. There was no issue found except for one thing:

Cloverleaf 6.2 host server keep stopping. We need to issue hciss -s h to start the host server of 6.2 now and then. Is there any logfile where we can look for issues?

Regards,

Varun

[Peter Heggie]

I’m glad you were able to resolve the issue.

I think there have been older posts on Clovertech that talked about problems with the Host Server. The only one I have had is where the Host Server or a related process filled up a log file and crashed, because it was trying to connect to sites that did not exist, over and over. You can check by looking in …/hci/cis6.x/integrator/server/server.ini – check the list of sites in this file, and make sure they are correct. Also check the log files in the …../server/logs – are any logs very large?

Good luck

Peter Heggie

[Author]

Replies