cancel
Showing results for 
Search instead for 
Did you mean: 

Log queue overflows in LiveCache

Former Member
0 Kudos

Dears,

In my system some problems are occurring with log queues.

The following message is presented in my system:

Log queue overflows: 70641, configured log queue pages: 2000 for each of 4 log queues

I think it is occurring because we are managing the Log with "Log automatic overwrite". This configuration is being used because we don´t have a lot of disk space to manage it as well.

Could you please advise me about this setup and a way to avoid this overflow?

My best Regards.

Al jey.

Accepted Solutions (0)

Answers (1)

Answers (1)

lbreddemann
Active Contributor
0 Kudos

Dears,

>

> In my system some problems are occurring with log queues.

> The following message is presented in my system:

> Log queue overflows: 70641, configured log queue pages: 2000 for each of 4 log queues

>

Ok, so what is the problem with that?

You use the liveCache quite heavily (e.g. make lots of commits/changes) and the storage where you put your log-volume to is not quick enough to save the data.

Either you fix the storage speed or you accept that your users might have to wait a bit longer for their application to save the data.

> I think it is occurring because we are managing the Log with "Log automatic overwrite". This configuration is being used because we don´t have a lot of disk space to manage it as well.

> Al jey.

These two things have no connection.

A log queue overflow simply means, that sessions have to wait until the log queue is saved to disk before they can put new entries to the log queue.

The log mode overwrite means: you don't care about data security and the ability to recover the liveCache from a backup.

It's a setting used for throw-away database instances where it doesn't matter to loose data.

It means that the database simply overwrites log entries that have not been backed up, when the log area gets filled.

regards,

Lars

Former Member
0 Kudos

I understood your reply. I found some performance problems in the storage, but the Unix administrator always closed your eyes to the problem. a lot of WP in LC waiting to W/R data in data area and log area due to this.

Your answer was what I wanted to hear.