Hi, Good Day !!

Webfocus Applciations were down when webfocus service and jboss applcation service were all up and running. Error message in applciation page: "There are no agents available on reporting server: EDASERVER"
There were no crashed or hung agents in server console. There were only 2 idle agents. On recycling the service, applications were back.

On analysis, we noticed multiple entries of error: "request by edapmon to notify agent is stopped". After which there were continous entry for "No free agents" until services were restarted.

Please help us identify the reason for this issue.

Thanks,
Mathavi

WebFOCUS 7.6.10 HF 5
JBOSS Application Server AS 5.1
Win 2003

The first messagfe you cite means you are just running out of agents. You can raise the maximum number of agents in the DEFAULT Service properties. The message "request by edapmon to notify agent is stopped" is issued when an agent has been used the number of times set in the "agent_refresh" property. It is refreshed the next time it is needed.

"The first messagfe you cite means you are just running out of agents."
But we noticed 2 idle agents in server console when the applciation were throwing this error. Confused on why weren't these idle agents been used ?

Thanks,
Mathavi

The server includes four pre-defined Data Services: DEFAULT, WC_DEFAULT, SCHED_DEFAULT, DFM_DEFAULT. You're running out of DEFAULT agents. The idle agents are likely WC_DEFAULT which handle Web Console requests.

If all your DEFAULT agents are in use, possible causes include:

• You have more users requesting reports simultaneously than the number of max agents you made available,
  
• You have users and report caster jobs using the same service (likely DEFAULT), resulting in more agents being needed than you thought,
  
• Report requests are taking long enough that an agent is still occupied when the next request comes in; this can be especially nasty with jobs run from report caster - those can quickly eat up your available agents.
  
• A database got stuck and is not sending back query results (this has happened to us when someone enabled tracing on our VMS/RDB data adapter)
  

Solutions include:

• Increase the number of max agents,
  
• Configure a special service for report caster jobs and configure report caster to use that,
  
• Optimize your procedures to run in less time
  
• If the problem originated from a repeated report caster job, run it less frequently
  
• Check whether your databases respond (especially take note of the one where you see the most agents in use on)
  

Hi, We had 2 idle agents under all categories - DEFAULT, WC_DEFAULt, DFM_DEFAULT, when we had this issue in the applicaiton page.

quote:

Originally posted by Alex:
The server includes four pre-defined Data Services: DEFAULT, WC_DEFAULT, SCHED_DEFAULT, DFM_DEFAULT. You're running out of DEFAULT agents. The idle agents are likely WC_DEFAULT which handle Web Console requests.

We have max agents set to 75 (DEFAULT)
Should we increase further more.

we do not have report caster jobs. an external scheduling tool has been used for schedules.

Our suspect in also on the application that wuld have consumed all agents.
But still confused that we noticed idle agents in server console when we saw this error (No free agents) in the application page.

Thanks,
Mathavi

quote:

Originally posted by Wep5622:
If all your DEFAULT agents are in use, possible causes include:

• You have more users requesting reports simultaneously than the number of max agents you made available,
  
• You have users and report caster jobs using the same service (likely DEFAULT), resulting in more agents being needed than you thought,
  
• Report requests are taking long enough that an agent is still occupied when the next request comes in; this can be especially nasty with jobs run from report caster - those can quickly eat up your available agents.
  
• A database got stuck and is not sending back query results (this has happened to us when someone enabled tracing on our VMS/RDB data adapter)
  

Solutions include:

• Increase the number of max agents,
  
• Configure a special service for report caster jobs and configure report caster to use that,
  
• Optimize your procedures to run in less time
  
• If the problem originated from a repeated report caster job, run it less frequently
  
• Check whether your databases respond (especially take note of the one where you see the most agents in use on)
  

If you want to get a notification from the Server when it runs out of agents, disk space or if an agent crashes you can add an SMTP host to the Servers configuration. Once that is done you can select the conditions you want to receive an email for. If you begin get emails notifying you of that agents ran out you can increase the number of agents. If you get crash emails then your emphasis would be on troubleshooting the cause of the crashes. You could also get "out of disk space" indicating that your disk is either filled up or filling up with temporary files.

Hi Alex,

Thank you for your post.

This is exactly the type of thing our shop needs to do.

We are a very small shop, and I am a developer; we do not have full-time WF admins, so with that I hope you do not mind a couple of follow up questions..

quote:

you can add an SMTP host to the Servers configuration.

By this do you mean that I would go to the WebFOCUS Administration Console and click 'Configuration' - 'Client Settings' - 'General'?

When I do this drill down I see a setting called 'WF_EMAIL_SERVER'.

Is 'WF_EMAIL_SERVER' the setting you are referring to?

I take it, I would enter into this setting something like the domain name of our physical server (SMTP server) that handles the sending of email, e.g., 'mail.mydomain.com'.

quote:

Once that is done you can select the conditions you want to receive an email for

Do you enter the conditions also within WF Admin Console?

If so, could you tell me where?

I've looked at the different menu options/setting pages, and couldn't see a slot for this.

Again, many thanks.

Dave

The SMTP host can be set in the Reporting Server Console. You can find all of the information you need in the Server Admin Manual (server_admin7702.pdf) downloadable from the IBI Tech Support site. SMTP configuration starts on page 247 and is followed by all of the events that can be reported. If you just go to the console the email configuration is found under Workspace --> Configuration

Thanks Alex! That was exactly what I needed to know.

I see now I was incorrect in looking at the WebFOCUS Administration Console.

I instead went to the Web Console and then under Workspace -> Configuration/Monitor -> E-mail SMTP Server/Event Routing, found what I was looking for.

Many thanks,
Dave

Thanks Alex and Dave.

I shall discuss with my team on applying the SMTP configuration to get the root cause of similar issue in future, to zero down to agent crash or agent unavailability.

But, is there a way we can get the root cause of this issue that occurred in the past (07/02)

Thanks,
Mathavi

If your edaprint log has no limit you might be able to scour it for clues. You can access these logs from the Server console. Good Luck.

Thanks for the response.
We have logs for last 12 days and missed to take back-up and lost our logs for the issue occurrence.

I have another question on the earlier post.

quote:

If you begin get emails notifying you of that agents ran out you can increase the number of agents.

Will this get effective immediately and agents are created for the application to use instantly.

Basically would like to know, if we can get rid of "out of agents" scenario without service restart.

Thanks again,
Mathavi

Changing the maximum number of agents does not require a restart. When you view the properties only those items marked with and asterisk require a restart.

Thanks Alex for all those detailed information.

quote:

Originally posted by Alex:
If you want to get a notification from the Server when it runs out of agents, disk space or if an agent crashes you can add an SMTP host to the Servers configuration. Once that is done you can select the conditions you want to receive an email for. If you begin get emails notifying you of that agents ran out you can increase the number of agents. If you get crash emails then your emphasis would be on troubleshooting the cause of the crashes. You could also get "out of disk space" indicating that your disk is either filled up or filling up with temporary files.

You will get hammered incessently if you do run out of agents, in 7.6.x. In 7.7.x, you may set a threshold of repeat messages. I have these notifications going to our pager and have had our carrier turn our pager off due to the amount of traffic coming through.

My quest to resolve the root issue (preventing running out of agents in the first place) has been very successful, though.

- ABT