lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rick Leir <rl...@leirtech.com>
Subject RE: How to store files larger than zNode limit
Date Wed, 14 Mar 2018 10:24:31 GMT
Markus, Atita
We set it higher too. 

When zk is recovering from a disconnected state it re-sends all the messages that it had been
trying to send while the machines were disconnected. Is this stored in a ' transaction log'
.tlog file? I am not clear on this. Zk also goes through the unsent messages when Solr starts
up, and startup can take a while longer.

With this in mind, it might make more sense to use zk for kbyte sized  blobs. But machines
are faster every year, so maybe Meg and Gig blobs will be appropriate. 
Cheers -- Rick

On March 13, 2018 5:56:56 PM EDT, Markus Jelsma <markus.jelsma@openindex.io> wrote:
>Hi - For now, the only option is to allow larger blobs via
>jute.maxbuffer (whatever jute means). Despite ZK being designed for kb
>sized blobs, Solr demands us to abuse it. I think there was a ticket
>for compression support, but that only stretches the limit.
>
>We are running ZK with 16 MB for maxbuffer. It holds the large
>dictionaries, it runs fine. 
>
>Regards,
>Markus
> 
>-----Original message-----
>> From:Atita Arora <atitaarora@gmail.com>
>> Sent: Tuesday 13th March 2018 22:38
>> To: solr-user@lucene.apache.org
>> Subject: How to store files larger than zNode limit
>> 
>> Hi ,
>> 
>> I have a use case supporting multiple clients and multiple languages
>in a
>> single application.
>> So , In order to improve the language support, we want to leverage
>the Solr
>> dictionary (userdict.txt) files as large as 10MB.
>> I understand that ZooKeeper's default zNode file size limit is 1MB.
>> I'm not sure sure if someone tried increasing it before and how does
>that
>> fares in terms of performance.
>> Looking at -
>https://zookeeper.apache.org/doc/r3.2.2/zookeeperAdmin.html
>> It states -
>> Unsafe Options
>> 
>> The following options can be useful, but be careful when you use
>them. The
>> risk of each is explained along with the explanation of what the
>variable
>> does.
>> jute.maxbuffer:
>> 
>> (Java system property:* jute.maxbuffer*)
>> 
>> This option can only be set as a Java system property. There is no
>> zookeeper prefix on it. It specifies the maximum size of the data
>that can
>> be stored in a znode. The default is 0xfffff, or just under 1M. If
>this
>> option is changed, the system property must be set on all servers and
>> clients otherwise problems will arise. This is really a sanity check.
>> ZooKeeper is designed to store data on the order of kilobytes in
>size.
>> I would appreciate if someone has any suggestions  on what are the
>best
>> practices for handling large config/dictionary files in ZK?
>> 
>> Thanks ,
>> Atita
>> 

-- 
Sorry for being brief. Alternate email is rickleir at yahoo dot com 
Mime
  • Unnamed multipart/alternative (inline, 7-Bit, 0 bytes)
View raw message