Many years ago, I accidentally ran Solr with the data dir on an NFS volume.
It was 100X slower.

wunder
Walter Underwood
wun...@wunderwood.org
http://observer.wunderwood.org/  (my blog)

> On Feb 26, 2020, at 2:42 PM, Vincenzo D'Amore <v.dam...@gmail.com> wrote:
> 
> Hi Massimiliano,
> 
> it’s not clear how much memory you have configured for your Solr instance. 
> And I would avoid an nfs mount for the datadir. 
> 
> Ciao,
> Vincenzo
> 
> --
> mobile: 3498513251
> skype: free.dev
> 
>> On 26 Feb 2020, at 19:44, Massimiliano Randazzo 
>> <massimiliano.randa...@gmail.com> wrote:
>> 
>> Il giorno mer 26 feb 2020 alle ore 19:30 Dario Rigolin <
>> dario.rigo...@comperio.it> ha scritto:
>> 
>>> You can avoid commit and leave solr do autocommit at certain times.
>>> Or use softcommit if you have search queries at the same time to answer.
>>> 550000 pages of 3500 words isn't a big deal for a solr server, what's the
>>> hardware configuration?
>> The solr instance runs on a server with the following configuration:
>> 12 core Intel(R) Xeon(R) CPU E5-2620 v3 @ 2.40GHz
>> 64GB Ram
>> solr's DataDir is on a volume of another server that I mounted via NFS (I
>> was thinking of moving the solr server to the server where the DataDir
>> resides even if it has lower characteristics 8 core Intel(R) Xeon(R) CPU
>>      E5506  @ 2.13GHz 24GB Ram)
>> 
>> What's you single solr document a single newspaper? a single page?
>> 
>> the single solr document refers to the single word of the document
>> 
>> 
>>> Do you have a solrcloud with 8 nodes? Or are you sending same document to 8
>>> single solr servers?
>>> I have 8 servers that process 550,000 newspapers and all of them write on
>> 1 solr server only
>> 
>> 
>>>> Il giorno mer 26 feb 2020 alle ore 19:22 Massimiliano Randazzo <
>>>> massimiliano.randa...@gmail.com> ha scritto:
>>>> Good morning
>>>> I have the following situation I have to index the OCR of about 550,000
>>>> pages of newspapers counting an average of 3,500 words per page and
>>> making
>>>> a document per word the records are many.
>>>> At the moment I have 1 instance of Solr and 8 servers that read and write
>>>> all on the same instance at the same time, at the beginning everything is
>>>> fine after a while when I add, delete or commit it gives me a TimeOut
>>> error
>>>> towards the solr server.
>>>> I suspect the problem is due to the fact that it is that I do many commit
>>>> operations of many docs at a time (practically if the newspaper is 30
>>> pages
>>>> I do 105,000 add and in the end I commit), if everyone does this and 8
>>>> servers within walking distance of each other I think this creates
>>> problems
>>>> for Solr.
>>>> What can I do to solve the problem?
>>>> Do I make a commi to each add?
>>>> Is it possible to configure the solr server to apply the add and delete
>>>> commands, and to commit it, the server autonomously supports the
>>> available
>>>> resources as it seems to do for the optmized command?
>>>> Reading the documentation I would have found this configuration to
>>>> implement but not if it solves my problem
>>>> <deletionPolicy class="solr.SolrDeletionPolicy">
>>>> <str name="maxCommitsToKeep">1</str>
>>>> <str name="maxOptimizedCommitsToKeep">0</str>
>>>> <str
>>> name="maxCommitAge">1DAY</str></deletionPolicy><infoStream>false</infoStream>
>>>> Thanks for your consideration
>>>> Massimiliano Randazzo
>>> --
>>> Dario Rigolin
>>> Comperio srl - CTO
>>> Mobile: +39 347 7232652 - Office: +39 0425 471482
>>> Skype: dario.rigolin
>> 
>> 
>> -- 
>> Massimiliano Randazzo
>> 
>> Analista Programmatore,
>> Sistemista Senior
>> Mobile +39 335 6488039
>> email: massimiliano.randa...@gmail.com
>> pec: massimiliano.randa...@pec.net

Reply via email to