hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jean-Daniel Cryans <jdcry...@apache.org>
Subject Re: Dead Servers
Date Thu, 25 Aug 2011 17:02:01 GMT
> First off, thanks for you response. 26 seconds seems a bit short to time
> outs o what are some more reasonable timeouts I should set?

Agreed, but then again it really depends on your configurations.
Setting a higher timeout would only hide the issue, fixing it is much
better.

>
> This is probably the root cause since my job was pretty hefty.

Do you have metrics installed on that cluster? Debugging this issues
while being blind ain't something I fancy doing, I guess it's the same
for others.

>
> Question about swapping...
> Make sure you don't swap, the JVM never behaves well under swapping
> Is this as simple setting
>
> sysctl -w vm.swappiness=5

Setting swappiness low is good, but if you overcommit your memory it
will still swap!

>
> I know its extremely situation dependent but what would be a recommended
> memory allocation to HBase... currently I have it set to 4G?

Depends on the available RAM? Check that other thread currently going
on on this mailing list about calculating memory assignment.

J-D

Mime
View raw message