One of all shard stopping, all shards stop

["hongkeun.yoo" <hu...@naver.com>]

Hi. 

I hava 3 shard solr cloud version 4.4.0 not replication.
<http://lucene.472066.n3.nabble.com/file/n4098015/ex1.png> 

for example, if one shard(leader) died for OOM, all shard is stop.

is it just the way that it is?
I want to find a option this problem.
I want to change 
if 1 shard died, remain shards request work nomally

thanks you.




--
View this message in context: http://lucene.472066.n3.nabble.com/One-of-all-shard-stopping-all-shards-stop-tp4098015.html
Sent from the Solr - User mailing list archive at Nabble.com.

Re: One of all shard stopping, all shards stop

[Erick Erickson <er...@gmail.com>]

I think if you set shards.tolerant=true you get information in the
return packet if a shard is completely down.

The other thing you can do is query the ZooKeeper cluster state
directly.

But I have to ask why you're not using a replica or two per shard.
That should provide automatic fail-over etc and make the necessity
of dealing with this case _much_ less frequent. Personally I'd put
more effort into making an always-up cluster than dealing with
when a single node goes down.

FWIW,
Erick


On Mon, Oct 28, 2013 at 8:10 AM, hongkeun.yoo <hu...@naver.com> wrote:

> Thanks for your reply. If one of server have stop and error, this
> option(distrib=false) is good work. Similarly option is
> "shards.tolerant=true". but I don't want to using this option. because the
> died server isn't show error message. only return not nothing data.
>
> I want to show error message at died server, the other way normal server
> work normally.
>
>
>
> --
> View this message in context:
> http://lucene.472066.n3.nabble.com/One-of-all-shard-stopping-all-shards-stop-tp4098015p4098053.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>

Re: One of all shard stopping, all shards stop

["hongkeun.yoo" <hu...@naver.com>]

Thanks for your reply. If one of server have stop and error, this
option(distrib=false) is good work. Similarly option is
"shards.tolerant=true". but I don't want to using this option. because the
died server isn't show error message. only return not nothing data.

I want to show error message at died server, the other way normal server
work normally.



--
View this message in context: http://lucene.472066.n3.nabble.com/One-of-all-shard-stopping-all-shards-stop-tp4098015p4098053.html
Sent from the Solr - User mailing list archive at Nabble.com.

Re: One of all shard stopping, all shards stop

["michael.boom" <my...@yahoo.com>]

When one of your shards dies, your index becomes incomplete. By default the
querying is distributed (on all shards - &distrib=true) and if one of them
(shard X) is down, then you get an error stating that there are "no servers
hosting shard X".

If the other shards are still up you can query them directly using
"&distrib=false" but in the resultset you will only have documents from that
shard. So you would have to query every active shard individually and then
merge the results yourself.
If I'm wrong please correct me.



-----
Thanks,
Michael
--
View this message in context: http://lucene.472066.n3.nabble.com/One-of-all-shard-stopping-all-shards-stop-tp4098015p4098024.html
Sent from the Solr - User mailing list archive at Nabble.com.