You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by Simon ELBAZ <se...@linagora.com> on 2019/05/02 14:37:14 UTC

Accidentaly removed SSTables of unneeded data

Hi,

I am running Cassandra v2.1 on a 3 node cluster.

/# yum list installed | grep cassa//
//cassandra21.noarch 2.1.12-1                         @datastax //
//cassandra21-tools.noarch 2.1.12-1                         @datastax /

Unfortunately, I accidentally removed the SSTables (using rm) (older 
than 10 days) of a table on the 3 nodes.

Running 'nodetool repair' on one of the 3 nodes returns error. Whereas, 
it does not on another.

I don't need to recover the lost data but I would like 'nodetool repair' 
not returning an error.

Thanks for any advice.

Simon


Re: Accidentaly removed SSTables of unneeded data

Posted by Simon ELBAZ <se...@linagora.com>.
Hi Shalom,

I've runned refresh as Nitan suggested without sstablescrub.

Then i tried drain/restart on the 3 nodes. The repair is now OK.

Thanks for your help

Simon

On 02/05/2019 16:58, shalom sagges wrote:
> Hi Simon,
>
> If you haven't did that already, try to drain and restart the node you 
> deleted the data from.
> Then run the repair again.
>
> Regards,
>
> On Thu, May 2, 2019 at 5:53 PM Simon ELBAZ <selbaz@linagora.com 
> <ma...@linagora.com>> wrote:
>
>     Hi,
>
>     I am running Cassandra v2.1 on a 3 node cluster.
>
>     /# yum list installed | grep cassa//
>     //cassandra21.noarch 2.1.12-1                         @datastax //
>     //cassandra21-tools.noarch 2.1.12-1                        
>     @datastax /
>
>     Unfortunately, I accidentally removed the SSTables (using rm)
>     (older than 10 days) of a table on the 3 nodes.
>
>     Running 'nodetool repair' on one of the 3 nodes returns error.
>     Whereas, it does not on another.
>
>     I don't need to recover the lost data but I would like 'nodetool
>     repair' not returning an error.
>
>     Thanks for any advice.
>
>     Simon
>

Re: Accidentaly removed SSTables of unneeded data

Posted by shalom sagges <sh...@gmail.com>.
Hi Simon,

If you haven't did that already, try to drain and restart the node you
deleted the data from.
Then run the repair again.

Regards,

On Thu, May 2, 2019 at 5:53 PM Simon ELBAZ <se...@linagora.com> wrote:

> Hi,
>
> I am running Cassandra v2.1 on a 3 node cluster.
>
> *# yum list installed | grep cassa*
> *cassandra21.noarch                2.1.12-1
> @datastax    *
> *cassandra21-tools.noarch          2.1.12-1
> @datastax   *
>
> Unfortunately, I accidentally removed the SSTables (using rm) (older than
> 10 days) of a table on the 3 nodes.
>
> Running 'nodetool repair' on one of the 3 nodes returns error. Whereas, it
> does not on another.
>
> I don't need to recover the lost data but I would like 'nodetool repair'
> not returning an error.
>
> Thanks for any advice.
>
> Simon
>

Re: Accidentaly removed SSTables of unneeded data

Posted by Nitan Kainth <ni...@gmail.com>.
You can run nodetool refresh and then sstablescrub to see if there is any
corruption.

On Thu, May 2, 2019 at 9:53 AM Simon ELBAZ <se...@linagora.com> wrote:

> Hi,
>
> I am running Cassandra v2.1 on a 3 node cluster.
>
> *# yum list installed | grep cassa*
> *cassandra21.noarch                2.1.12-1
> @datastax    *
> *cassandra21-tools.noarch          2.1.12-1
> @datastax   *
>
> Unfortunately, I accidentally removed the SSTables (using rm) (older than
> 10 days) of a table on the 3 nodes.
>
> Running 'nodetool repair' on one of the 3 nodes returns error. Whereas, it
> does not on another.
>
> I don't need to recover the lost data but I would like 'nodetool repair'
> not returning an error.
>
> Thanks for any advice.
>
> Simon
>