Please help urgently: QV Log_files growing way too big

[datanibbler]

Hi,

please help me: One of the QV Log_Files - the Distribution_Service logs, the ones labeled "root_YYYYMMDD.txt", are growing way too big. I'm not trained in administration of any kind. At last I have access to the logs, so here's some welcome 😉

Precisely, three of them - since July 19, 2014 - are extremely big - 32 MB on July 19, 43 MB on July 20 and 15MB on July 21. The ones before have a size of 100, max. 500KB.

We have just done an update to the most current version with the help of an external consultant.

I have already opened a ticket with our partner CP who in turn will open a ticket with QlikTech. In the biggest of the logfiles, there are some lines which, if I'm not mistaken, repeat over and over all day:

20/07/2014 00:05:02.3019625 Error PurgeThread encountered exception.. Exception=System.ArgumentException: An item with the same key has already been added.

   at System.Collections.Generic.Dictionary`2.Insert(TKey key, TValue value, Boolean add)

   at System.Collections.Generic.Dictionary`2.Add(TKey key, TValue value)

   at QDSMain.TaskStore.AddRange(IEnumerable`1 tasks)

   at QDSMain.Workorder.get_TaskAndTempTaskList()

   at QDSMain.TaskLogIndex.RemoveAllWithoutLogFile()

   at QDSMain.FileSystemHelper.RunPurgeMethods()

As far as I can see at first glance, it goes on like this all day - always the same lines.

Can you make anything of this?

Thanks a lot!

Best regards,

DataNibbler

[#2]

Did you restarted the Services in the last time? Could be an communication issue between Management Service and Distribution Service.

[#3]

What version are you running? I remember a similar issue a while back when on v9, had to set up an automated service restart once every 2 weeks to counter the problem

[datanibbler]

Hi,

@ Daniel

Yes - we did restart just about everything you possibly can restart 😉 And we certainly did during the update which we installed as part of our "crisis management" these days.

@ Joseph

We are now running vrs. 11, SR7. We have had to restart the DistributionService once in a while some months back, we thought about setting up an automated task to do that at the time, but for some reason which I don't remember that plan remained just a plan. During the last weeks, I don't remember anything of the sort.

A possible short-term solution would be to just delete the log_files, wouldn't it? - I think they would just be re-created, no damage would be done. But that would of course not be a long-term solution as long as we don't know why this happens.

P.S.: I just notice something now - the biggest of those root.txt files are on July 19, 20 and 21. One on July 22 is just 51KB in size - there are none after that. From that, I guess that the service has possibly balanced itself - still it seems necessary to know why this happened.

[#5]

Yea I had to trim down the large logs rather than delete (so we still had a full log history day by day) but deleting should be okay too.

[datanibbler]

Hi Joseph,

nope - I cannot delete the logs 😉 Apparently, the colleague has only granted me RO access to the log_files.

That way I'm not so sure the system_monitor or Governance_Dashboard will work at all?

Can you advise me on this?

Thanks a lot!

[#7]

Both are great tools to see what is going on. The System Monitor is more deep when analysing all type of Logs. The Governance Dashboard is great to analyse your Applications and Data.

[datanibbler]

Hi Daniel,

thanks for this clarification! I guess the system_monitor is more like what I need then.

- Knowing which apps use what data is m.o.l. irrelevant - or, rather, known - as I have developed most of them
   myself.
- Knowing which of my apps are used how intensively would be interesting to know, but not really anything I should
   be concerned about - for all I know, some of them
   are not widely used because, since we don't have that many licenses, I have assigned licenses to central team-
    accounts - every team has one - but not all team_members know the logon_data for this 😉
 

[datanibbler]

Hi,

can anyone tell me positively whether the system_monitor is going to work with RO access to the log_files? If not, I would have to ask the colleague from IT to grant me RW access.

Thanks a lot!

[#10]

This should work, as it just read the Logs. Or use the Server to reload, as the Service User will have full access to write the Logs.