You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by Furkan KAMACI <fu...@gmail.com> on 2013/04/30 10:51:37 UTC

SolrCloud terminology - Why leader is listed as slave and replica is listed as master at admin gui?

When I look at admin gui I see that for a leader:

Replication (Slave) Version Gen Size
Master: 1367309548534 84 779.87 MB
Slave:  1367307649512 82 784.44 MB

and that for a replica:

Replication (Master) Version Gen Size
Master: 1367309548534 84 779.87 MB

isn't it confusing leader is a slave and replica is a master or is it true
for SolrCloud terminology?

Re: SolrCloud terminology - Why leader is listed as slave and replica is listed as master at admin gui?

Posted by Shawn Heisey <so...@elyograg.org>.
On 4/30/2013 8:16 AM, Shawn Heisey wrote:
> Master info gets listed as soon as the replication handler has started.
>  If that core has ever been used as a replication target since the last
> restart, then it will show the slave information as well.

After thinking about it. I'm not 100% sure that what I said was correct.
 I don't have time right now to do any experimentation.

Thanks,
Shawn



Re: SolrCloud terminology - Why leader is listed as slave and replica is listed as master at admin gui?

Posted by Shawn Heisey <so...@elyograg.org>.
On 4/30/2013 2:51 AM, Furkan KAMACI wrote:
> When I look at admin gui I see that for a leader:
> 
> Replication (Slave) Version Gen Size
> Master: 1367309548534 84 779.87 MB
> Slave:  1367307649512 82 784.44 MB
> 
> and that for a replica:
> 
> Replication (Master) Version Gen Size
> Master: 1367309548534 84 779.87 MB
> 
> isn't it confusing leader is a slave and replica is a master or is it true
> for SolrCloud terminology?

Master info gets listed as soon as the replication handler has started.
 If that core has ever been used as a replication target since the last
restart, then it will show the slave information as well.

On the last restart, your current leader probably replicated itself from
the other replica, which would have been leader at the time.  Your other
replica has never (since the last restart) been used as a replication
target, it has only ever kept itself in sync by doing SolrCloud
distributed indexing.

There is a bug in the replication info display, fixed in 4.3.  The bug
info says it affects version 4.2, but I would not be surprised to learn
that it affects all previous 4.x releases as well.  This doesn't cause
any actual problem with replication, just the confusing information on
the slave:

https://issues.apache.org/jira/browse/SOLR-4661

The 4.3 release is on release candidate 4, hoping for a release just
after Lucene Revolution 2013 ends.

Thanks,
Shawn