lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tomás Fernández Löbbe (JIRA) <j...@apache.org>
Subject [jira] [Updated] (SOLR-12213) Consider removing the collectionPropsNotifications thread from ZkStateReader
Date Wed, 11 Apr 2018 21:00:00 GMT

     [ https://issues.apache.org/jira/browse/SOLR-12213?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Tomás Fernández Löbbe updated SOLR-12213:
-----------------------------------------
    Affects Version/s: master (8.0)

> Consider removing the collectionPropsNotifications thread from ZkStateReader
> ----------------------------------------------------------------------------
>
>                 Key: SOLR-12213
>                 URL: https://issues.apache.org/jira/browse/SOLR-12213
>             Project: Solr
>          Issue Type: Improvement
>      Security Level: Public(Default Security Level. Issues are Public) 
>    Affects Versions: 7.4, master (8.0)
>            Reporter: Tomás Fernández Löbbe
>            Priority: Minor
>
> From SOLR-12172:
> [~shalinmangar]
> {quote}
> Tomás Fernández Löbbe – I don't think we should introduce another thread(pool) just
for this feature. We can use a method similar to updateWatchedCollection which checks if the
new znode version is greater than the old one. This ensures that we replace the old collection
props only if the new one is actually newer.
> {quote}
> [~tomasflobbe]
> {quote}
> Thanks for the review Shalin Shekhar Mangar. I thought about doing something like that,
but decided not to since it requires to keep something like a map with collection -> version
and handling it made the code more complex. I'll put up a patch, maybe it's still better to
go that route anyway
> {quote}
> [~tomasflobbe]
> {quote}
> Shalin Shekhar Mangar, just by keeping the synchronization I added to refreshAndWatch
in the previous commit we can guarantee that we won't be setting the collection property map
to an older value, however, I don't think we can guarantee that the notifications to watchers
won't be out of order without using the single thread executor. Are you suggesting that we
go that way anyway?
> {quote}



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message