ndb-log-bin is ON by default. I would not say this must be turned off by
default. I am running 1 particular benchmark, update of a non-indexed
column, in large batch sets (64, 128, 256 threads).
Update of an indexed column runs fine with no issues and I was able to
insert a million rows very rapidly as well with no issues. Which makes
this even more perplexing. :) (well, the million insert is just 1
thread..)
Yes, I do see that message in the manager's log. No idea what they mean
either because those messages don't seem to coincide with the overloaded
kernel error I'm getting.
-Matthew
> -----Original Message-----
> From: Robert Klikics [mailto:server@stripped]
> Sent: Wednesday, November 11, 2009 12:47 PM
> To: cluster@stripped
> Subject: Re: Send Buffers overloaded in NDB kernel
>
> Hi Metthew,
>
> > Pretty much have to just turn off ndb-log-bin (on each mysqld node
> > connected to the cluster) so that tables that are ENGINE=NDB don't
> get
> > logged in the binary log.
>
> does that mean that ndb-log-bin "ON" is the default behaviour in MySql
> cluster? Should this be changed to "NO" by default?
> And, if the normal "log-bin" is disables, it seems that "ndb-log-bin"
> is also disabled.
>
> Did you find log-entrys like this on your ndb_*_cluster.log:
>
> Event buffer status: used=822KB(45%) alloc=1801KB(0%) max=0B
> apply_epoch=7239717/9 latest_epoch=7239717/9
>
> ?
>
> I don't know what this means. Any ideas?
>
> Regards,
> Robert
> --
> MySQL Cluster Mailing List
> For list archives: http://lists.mysql.com/cluster
> To unsubscribe:
> http://lists.mysql.com/cluster?unsub=1
