manifoldcf-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Karl Wright (JIRA)" <>
Subject [jira] [Commented] (CONNECTORS-1367) Alfresco Connector deadlock with postgresql
Date Mon, 16 Jan 2017 12:19:26 GMT


Karl Wright commented on CONNECTORS-1367:

[~chalitha.perera]: I've looked at this in some detail now.  The Alfresco Webscript connector
uses a specific connector model which makes continuous crawling not possible.  That is why
the selection doesn't appear in the UI.

The job lifecycle for this connector is that there is indeed only ONE seeding invocation per
job execution.  That seeding execution is not completing, for some reason, and that is clearly
the source of the problem here.

Can you look at the Alfresco server logs to see whether there's an exception dumped there
that might explain the inability to respond to the seeding request?  I expect that you will
find an "out of memory" exception there if you look.  If that is the case, that basically
means that the seeding request must be "paged", because there are too many documents.  Maurizio
Pillitu is the person we'll need to look at how to do that.

Please confirm, first, that that is what is happening.  Thanks!

> Alfresco Connector deadlock with postgresql
> -------------------------------------------
>                 Key: CONNECTORS-1367
>                 URL:
>             Project: ManifoldCF
>          Issue Type: Bug
>          Components: Alfresco webscript connector
>    Affects Versions: ManifoldCF 2.5
>            Reporter: Chalitha Perera
>            Assignee: Karl Wright
>         Attachments: jstack5.out
> Alfresco connector gets stuck when processing documents around 35,000 and postgres manifold
database in idle state and after days gets deadlock message 
> "
> 2016-12-21 10:54:17.946 UTC >ERROR:  deadlock detected
> < 2016-12-21 10:54:17.946 UTC >DETAIL:  Process 28072 waits for ShareLock on transaction
432231; blocked by process 28038.
>     Process 28038 waits for ShareLock on transaction 432227; blocked by process 28072.
>     Process 28072: UPDATE jobqueue SET failtime=NULL,checktime=$1,failcount=NULL,checkaction=$2
WHERE jobid=$3 AND  (status=$4 OR status=$5)
>     Process 28038: SELECT needpriority FROM jobqueue WHERE id=$1 FOR UPDATE
> < 2016-12-21 10:54:17.946 UTC >HINT:  See server log for query details."
> On separate note, ran ant load-pg and it was completed successfully.

This message was sent by Atlassian JIRA

View raw message