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 Erick Erickson <er...@gmail.com> on 2013/08/01 15:22:09 UTC

Re: SolrCloud - Replica 'down'. How to get it back as 'active'? - Solr 4.3.0

Soft commits have no influence on the transaction log. You must do a hard
commit (openSearcher=true|false doesn't matter) to truncate the current
tlog and start a new one.

You're replaying all of your index insertions for the last 14 hours so it
might take quite a long time.

Best
Erick


On Wed, Jul 31, 2013 at 11:25 AM, Jeroen Steggink <je...@stegg-inc.com>wrote:

> Thanks Anshum,
>
> autoSoftCommit was alread set to 1000ms, but I changed the autoCommit to 3
> minutes.
>
> I'll wait for it to come back. The index contains about 200.000 documents
> and the last commit was 14 hours ago. So I wonder how long it will take.
> I would have thought it would be back up already.
>
>
> On 31-7-2013 16:40, Anshum Gupta wrote:
>
>> It perhaps is just replaying the transaction logs and coming up. Wait for
>> it is what I'd say.
>> The admin UI as of now doesn't show replaying of transaction log as
>> 'recovering', it does so only during peer sync.
>>
>> Also, you may want to add autoSoftCommit and increase the autoCommit to a
>> few minutes.
>>
>>
>> On Wed, Jul 31, 2013 at 7:55 PM, Jeroen Steggink <jeroen@stegg-inc.com
>> >wrote:
>>
>>  Hi,
>>>
>>> After the following error, one of the replicas of the leader went down.
>>> "Error opening new searcher. exceeded limit of maxWarmingSearchers=2, try
>>> again later."
>>> I increased the autoCommit time to 5000ms and restarted Solr.
>>>
>>> However, the status is still set to "down".
>>> How do I get it back to "active"?
>>>
>>> Regards,
>>> Jeroen
>>>
>>>
>>>
>>>
>>
>
>