Re: REDO log files overloaded

Pekka Nousiainen wrote:
> It's a temporary error. You have to re-run the transaction.
> This is how NDB, by design, deals with temporary errors.
> Maybe this is your long term plan?
>
> Otherwise, run your inserts and updates a little bit
> slower. You could also increase RedoBuffer but this does
> not really help if you constantly put in more data than
> redo logging can handle.

The long term solution is to reconfigure the cluster. We added a new application to the cluster yesterday and it seems we have overloaded the redo logs in the process. I just need to do all the math again for our new level of throughput. We were successful in testing, but testing always seems to never be quite the same as the real thing.

 > The "at all" part seems strange.

Yes, that was the strange part. I shut off all access to the cluster for an hour, came back and tried an update and got the same REDO error again. Maybe it needed more than an hour.

-- 

Brian Moon
Senior Developer
------------------------------
http://dealnews.com/
It's good to be cheap =)

-- 
MySQL Cluster Mailing List
For list archives: 
http://lists.mysql.com/cluster
To unsubscribe:    
http://lists.mysql.com/cluster?unsub=lists@pantek.com