You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by Roger Warner <rw...@pandora.com> on 2017/11/30 01:42:06 UTC
Nodetool repair on read only cluster
What would running a repair on a cluster do when there are no deletes nor have there ever been? I have no deletes yet on my data. Yet running a repair took over 9 hours on a 5 node cluster?
Roger?
Re: Nodetool repair on read only cluster
Posted by Jeff Jirsa <jj...@gmail.com>.
Over time the various nodes likely got slightly out of sync - dropped mutations primarily, during Long GC pauses or maybe network failures
In that case, repair will make all of the data match - how long it takes depends on size of data (more data takes longer to validate), size of your partitions (big partitions are more work to repair), and how you invoke repair
--
Jeff Jirsa
> On Nov 29, 2017, at 5:42 PM, Roger Warner <rw...@pandora.com> wrote:
>
>
> What would running a repair on a cluster do when there are no deletes nor have there ever been? I have no deletes yet on my data. Yet running a repair took over 9 hours on a 5 node cluster?
>
> Roger?
Re: Nodetool repair on read only cluster
Posted by "@Nandan@" <na...@gmail.com>.
Hi Roger,
As you provide incomplete information which is so tough to analyse .
But if you like to refer then please check below JIRA link to check out is
it useful or not. ?
https://issues.apache.org/jira/browse/CASSANDRA-6616
Thanks.
On Thu, Nov 30, 2017 at 9:42 AM, Roger Warner <rw...@pandora.com> wrote:
>
>
> What would running a repair on a cluster do when there are no deletes nor
> have there ever been? I have no deletes yet on my data. Yet running a
> repair took over 9 hours on a 5 node cluster?
>
>
>
> Roger?
>