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 Chuck Reynolds <cr...@ancestry.com> on 2019/05/24 13:16:26 UTC
Restore is creating the cluster incorrectly
I have 4 instances of Solr running on 3 servers with a replication factor of 3.
When I execute the command to do the restore to a new TEST cluster it create each master with the same IP address and port but all subsequent replicas are create correctly. That means that the other 3 instances of solr running on server 10.157.120.75 are not being used.
Any help would be appreciated
[cid:image001.png@01D51200.98EE1F20]
Re: Restore is creating the cluster incorrectly
Posted by Chuck Reynolds <cr...@ancestry.com>.
We are running Solr 6.6
Sent from my iPhone
> On May 25, 2019, at 12:45 PM, Mikhail Khludnev <mk...@apache.org> wrote:
>
> Hello, Chuck.
>
> Which version do you run?
> Can't you encounter https://urldefense.proofpoint.com/v2/url?u=https-3A__issues.apache.org_jira_browse_SOLR-2D9527&d=DwIBaQ&c=kKqjBR9KKWaWpMhASkPbOg&r=J-2s3b-3-OTA0o6bGDhJXAQlB5Y3s4rOUxlh_78DJl0&m=ce8IRTn9v0DazVrIgbbAzskEtjbmL79RLg0mB6wncrc&s=hBytAZMffpfESuSDUnIxDTFpsoc62YAxXvxCf1r2bAs&e= ?
>
> On Fri, May 24, 2019 at 7:52 PM Chuck Reynolds <cr...@ancestry.com>
> wrote:
>
>>
>>
>> I have 4 instances of Solr running on 3 servers with a replication factor
>> of 3. They are using ports 10001 -10004
>>
>>
>>
>> Server 1 10.xxx.xxx.75
>>
>> Server 2 10.xxx.xxx.220
>>
>> Server 3 10.xxx.xxx.245
>>
>>
>>
>> When I execute the command to do the restore to a new cluster it create
>> each master with the same IP address and port but all subsequent replicas
>> are create correctly. Why does it create the master using the same server
>> and port?
>>
>>
>>
>> This means that one of the 4 solr instances on server 10.xxx.xxx.75 is
>> managing 4 shards of data while some of the other instances are not
>> managing any shards of data.
>>
>>
>>
>> I now that running more than one instances of Solr on the same server is
>> not standard but I can setup this same cluster with 4 instances of Solr
>> running on a single server and the create collection command is smart
>> enough to figure it out.
>>
>>
>>
>> *Shard1*
>>
>> 10.xxx.xxx.75:10002 * master*
>>
>> 10.xxx.xxx.220:10001 repl
>>
>> 10.xxx.xxx.220:10003 repl
>>
>>
>>
>> *Shard2*
>>
>> 10.xxx.xxx.75:10002 * master*
>>
>> 10.xxx.xxx.245:10003 repl
>>
>> 10.xxx.xxx.75:10004 repl
>>
>>
>>
>> *Shard3*
>>
>> 10.xxx.xxx.75:10002 * master*
>>
>> 10.xxx.xxx.245:10004 repl
>>
>> 10.xxx.xxx.75:10003 repl
>>
>>
>>
>> *Shard4*
>>
>> 10.xxx.xxx.75:10002 * master*
>>
>> 10.xxx.xxx.245:10001 repl
>>
>> 10.xxx.xxx.220:10002 repl
>>
>>
>>
>>
>>
>>
>>
>> Any help would be appreciated
>>
>>
>>
>> [image: cid:image001.png@01D51200.98EE1F20]
>>
>>
>>
>
>
> --
> Sincerely yours
> Mikhail Khludnev
Re: Restore is creating the cluster incorrectly
Posted by Mikhail Khludnev <mk...@apache.org>.
Hello, Chuck.
Which version do you run?
Can't you encounter https://issues.apache.org/jira/browse/SOLR-9527 ?
On Fri, May 24, 2019 at 7:52 PM Chuck Reynolds <cr...@ancestry.com>
wrote:
>
>
> I have 4 instances of Solr running on 3 servers with a replication factor
> of 3. They are using ports 10001 -10004
>
>
>
> Server 1 10.xxx.xxx.75
>
> Server 2 10.xxx.xxx.220
>
> Server 3 10.xxx.xxx.245
>
>
>
> When I execute the command to do the restore to a new cluster it create
> each master with the same IP address and port but all subsequent replicas
> are create correctly. Why does it create the master using the same server
> and port?
>
>
>
> This means that one of the 4 solr instances on server 10.xxx.xxx.75 is
> managing 4 shards of data while some of the other instances are not
> managing any shards of data.
>
>
>
> I now that running more than one instances of Solr on the same server is
> not standard but I can setup this same cluster with 4 instances of Solr
> running on a single server and the create collection command is smart
> enough to figure it out.
>
>
>
> *Shard1*
>
> 10.xxx.xxx.75:10002 * master*
>
> 10.xxx.xxx.220:10001 repl
>
> 10.xxx.xxx.220:10003 repl
>
>
>
> *Shard2*
>
> 10.xxx.xxx.75:10002 * master*
>
> 10.xxx.xxx.245:10003 repl
>
> 10.xxx.xxx.75:10004 repl
>
>
>
> *Shard3*
>
> 10.xxx.xxx.75:10002 * master*
>
> 10.xxx.xxx.245:10004 repl
>
> 10.xxx.xxx.75:10003 repl
>
>
>
> *Shard4*
>
> 10.xxx.xxx.75:10002 * master*
>
> 10.xxx.xxx.245:10001 repl
>
> 10.xxx.xxx.220:10002 repl
>
>
>
>
>
>
>
> Any help would be appreciated
>
>
>
> [image: cid:image001.png@01D51200.98EE1F20]
>
>
>
--
Sincerely yours
Mikhail Khludnev
Restore is creating the cluster incorrectly
Posted by Chuck Reynolds <cr...@ancestry.com>.
I have 4 instances of Solr running on 3 servers with a replication factor of 3. They are using ports 10001 -10004
Server 1 10.xxx.xxx.75
Server 2 10.xxx.xxx.220
Server 3 10.xxx.xxx.245
When I execute the command to do the restore to a new cluster it create each master with the same IP address and port but all subsequent replicas are create correctly. Why does it create the master using the same server and port?
This means that one of the 4 solr instances on server 10.xxx.xxx.75 is managing 4 shards of data while some of the other instances are not managing any shards of data.
I now that running more than one instances of Solr on the same server is not standard but I can setup this same cluster with 4 instances of Solr running on a single server and the create collection command is smart enough to figure it out.
Shard1
10.xxx.xxx.75:10002 master
10.xxx.xxx.220:10001 repl
10.xxx.xxx.220:10003 repl
Shard2
10.xxx.xxx.75:10002 master
10.xxx.xxx.245:10003 repl
10.xxx.xxx.75:10004 repl
Shard3
10.xxx.xxx.75:10002 master
10.xxx.xxx.245:10004 repl
10.xxx.xxx.75:10003 repl
Shard4
10.xxx.xxx.75:10002 master
10.xxx.xxx.245:10001 repl
10.xxx.xxx.220:10002 repl
Any help would be appreciated
[cid:image001.png@01D51200.98EE1F20]
Re: Restore is creating the cluster incorrectly
Posted by Shawn Heisey <ap...@elyograg.org>.
On 5/24/2019 7:16 AM, Chuck Reynolds wrote:
> I have 4 instances of Solr running on 3 servers with a replication
> factor of 3.
>
> When I execute the command to do the restore to a new *TEST* cluster it
> create each master with the same IP address and port but all subsequent
> replicas are create correctly. That means that the other 3 instances of
> solr running on server 10.157.120.75 are not being used.
>
> Any help would be appreciated
The picture that you tried to include here did not make it to the list.
Email attachments rarely do.
Can you be more descriptive about the problem? What actions are you
taking, what are you expecting, and what actually happens?
Thanks,
Shawn