Hi,

I saw a post for this sort of problem on the forum in the following link All EDA Agents being used up by one report.

There was little feedback on this, has anyone since had a similar problem of a single report using multiple agents. Yesterday a user ran an adhoc report and I watched a new agent 'launch' every 3 minutes (exactly 3 min apart ) for the same report. Needless to say I stopped it at 5 agents as the server was being severely hammered.

Does anyone have any idea what might cause this, or is this something to log with IBI?

Thanks,

Jinx.

This message has been edited. Last edited by: Kerry,

Funny you should bring this up now, as I have just had an issue just like this.

I had a job that ended up in an infinite loop, with code in between, and looks like it spawned 3 other agents. These were hours apart.

Weird!
We had a thought yesterday that maybe it had something to do with our web front end calling the report multiple times (as c# and asp can do - though they would only call the report twice in certain conditions, like a duplicated on-click event) but the fact that it spawned a new agent every 3 minutes on the nose says it is something within webfocus.
I suppose it's down to logging a call with ibi. I just cannot replicate this problem, the user who was affected seems to have no problems now, so it's totally intermittent.

Jinx.

If it helps, I have seen similar before, and in my case its always been a long running report.

Your problem may be very different.

I would suggest, if it happens again, to gather any info possible from the WF console, to see what master its working on, you may notice it repeating the same master if its looping. Also if you have access to the EDATEMP directory, look at the files

Jinx,

I have had the exact same problem that you are experiencing although we could never solve it. It started out like with you where it was intermittent and only happened with one or two users (me being one of those, but not the other developers). But it progressively got worse and started affecting more and more users throughout the company. The only common thread that I was ever able to figure out was that it would happen on any report longer then 3 minutes.

This is my post from a couple years ago

http://forums.informationbuild...141067992#8141067992

So this is not a new problem and it will only get worse. I found that this only happened in foreground reports and never happened in defered reports. We had to start telling our users to fire off most things defered because we couldn't solve the issue.

This did eventually go away for us, but not until we outsourced our servers to a third party. My suggestion is to get both your network people involved and IBI. I don't know where the issue lies, but my guess is that it is somewhere with the interaction between the 2.

I hope you can find a resolution where we couldn't.

Eric

Thanks for all of the feedback. We are going to look into it some more and try and identify conditions under which it may occur, and depnding on how often it happens (after checking the report run logs) possibly open a case with IBI.
It's a worrying bug since a user could run a report at the weekend/evening, and the database would basically be hammered, when there is little to no IT support available. *ponder* Not good.
We'll also stop blaming the users for running the same report multiple times! ;-)

Thanks,

Jinx.

Jinx,

When this was occurring to us, it wasn't as large a problem as it seems like it will be for you because we limit our users to 3 foreground agents and 2 defered agents. Maybe that is something you might want to consider doing as well. Limit the amount of foreground agents that your users can run to 3 or 5 so that the problem can be contained and your systems don't get pegged so bad no one can get any reports.

Hi,

I didn't know that one could do that! So we have several hundred users at any one time, can I limit each user to just 1 or 2 agents? If that is the case it would be great either way - for those times when they do click the submit button multiple times! Also with this bug, it would not be such a major issue since a couple of agents launching won't totally kill it the database server.

Thanks,

Jinx.

Jinx,

If you go to your console and click on workspace > Configuration you will see all of your data services. We have it set up so that our default service will allow someone to run 3 agents, Our defered service (DFM_DEFAULT) allows 2 agents and then the rest that a user submits gets put into a queue and fired off as agents are made available. We have a service specifically for report caster reports and then we have what we called a dashboard service. In our dashboard service we opened it up completly, unlimited agents. We did this because all the queries that we put through there are fast queries that take no longer then 30 seconds (thats fast right now in our world). People are in and out of our databases and not causing problems. Also as we started building bigger dashboard applications that needed to fire off more then 3 agents simultaneously. Creating that extra service solved it for us.


One thing to keep in mind though is that when a user utilizes all of their agents, they will receive an error message saying that they are out of agents. I would just make sure you are aware of what the error message says and ok with your users being presented with it. You probably can go into the html for the error message and change it if you want, but I've never tried looking for the file to do that.