Re: Change way of logging OOM and InterruptedException in WorkerThread

2022-03-15 Thread Karl Wright
Yes, you can do that.   Thanks!
Karl


On Tue, Mar 15, 2022 at 2:19 PM Julien Massiera <
julien.massi...@francelabs.com> wrote:

> Then I'll create a ticket to just add a date to the system.err.println
> output of the OOM, is it ok ?
>
> -Message d'origine-
> De : Karl Wright 
> Envoyé : mardi 15 mars 2022 19:10
> À : dev 
> Objet : Re: Change way of logging OOM and InterruptedException in
> WorkerThread
>
> (1) InterruptedExceptions are thrown during normal shutdown and should
> generally not be logged.  This is by design.
> (2) OOM conditions can prevent logging from working.  So if you want to
> log those, fine, but please retain the System.err output.
>
> Karl
>
>
> On Tue, Mar 15, 2022 at 1:51 PM Julien Massiera <
> julien.massi...@francelabs.com> wrote:
>
> > Hi,
> >
> >
> >
> > In the WorkerThread class, OOM are logged using System.err (l849) and
> > therefore cannot be formated, but more importantly, the timestamp is
> > lost, which can be very problematic to investigate issues. Also,
> > InterruptedExceptions are not logged at all (l844) !
> >
> > Can we use Logging.threads.fatal for OOM and  Logging.threads.error
> > for InterruptedExceptions instead ?
> >
> >
> >
> > If it is ok, I can create a ticket and make the changes myself.
> >
> >
> >
> > Regards,
> >
> > Julien
> >
> >
> >
> >
> >
> >
>
>


RE: Change way of logging OOM and InterruptedException in WorkerThread

2022-03-15 Thread Julien Massiera
Then I'll create a ticket to just add a date to the system.err.println output 
of the OOM, is it ok ? 

-Message d'origine-
De : Karl Wright  
Envoyé : mardi 15 mars 2022 19:10
À : dev 
Objet : Re: Change way of logging OOM and InterruptedException in WorkerThread

(1) InterruptedExceptions are thrown during normal shutdown and should 
generally not be logged.  This is by design.
(2) OOM conditions can prevent logging from working.  So if you want to log 
those, fine, but please retain the System.err output.

Karl


On Tue, Mar 15, 2022 at 1:51 PM Julien Massiera < 
julien.massi...@francelabs.com> wrote:

> Hi,
>
>
>
> In the WorkerThread class, OOM are logged using System.err (l849) and 
> therefore cannot be formated, but more importantly, the timestamp is 
> lost, which can be very problematic to investigate issues. Also, 
> InterruptedExceptions are not logged at all (l844) !
>
> Can we use Logging.threads.fatal for OOM and  Logging.threads.error 
> for InterruptedExceptions instead ?
>
>
>
> If it is ok, I can create a ticket and make the changes myself.
>
>
>
> Regards,
>
> Julien
>
>
>
>
>
>



Re: Change way of logging OOM and InterruptedException in WorkerThread

2022-03-15 Thread Karl Wright
(1) InterruptedExceptions are thrown during normal shutdown and should
generally not be logged.  This is by design.
(2) OOM conditions can prevent logging from working.  So if you want to log
those, fine, but please retain the System.err output.

Karl


On Tue, Mar 15, 2022 at 1:51 PM Julien Massiera <
julien.massi...@francelabs.com> wrote:

> Hi,
>
>
>
> In the WorkerThread class, OOM are logged using System.err (l849) and
> therefore cannot be formated, but more importantly, the timestamp is lost,
> which can be very problematic to investigate issues. Also,
> InterruptedExceptions are not logged at all (l844) !
>
> Can we use Logging.threads.fatal for OOM and  Logging.threads.error for
> InterruptedExceptions instead ?
>
>
>
> If it is ok, I can create a ticket and make the changes myself.
>
>
>
> Regards,
>
> Julien
>
>
>
>
>
>


Change way of logging OOM and InterruptedException in WorkerThread

2022-03-15 Thread Julien Massiera
Hi,

 

In the WorkerThread class, OOM are logged using System.err (l849) and
therefore cannot be formated, but more importantly, the timestamp is lost,
which can be very problematic to investigate issues. Also,
InterruptedExceptions are not logged at all (l844) !

Can we use Logging.threads.fatal for OOM and  Logging.threads.error for
InterruptedExceptions instead ?

 

If it is ok, I can create a ticket and make the changes myself.

 

Regards,

Julien 

 

 



[jira] [Resolved] (CONNECTORS-1700) TikaServiceRmeta: Add options to filter out metadata based on size

2022-03-15 Thread Julien Massiera (Jira)


 [ 
https://issues.apache.org/jira/browse/CONNECTORS-1700?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Julien Massiera resolved CONNECTORS-1700.
-
Fix Version/s: ManifoldCF next
   Resolution: Fixed

r1898949

> TikaServiceRmeta: Add options to filter out metadata based on size
> --
>
> Key: CONNECTORS-1700
> URL: https://issues.apache.org/jira/browse/CONNECTORS-1700
> Project: ManifoldCF
>  Issue Type: Improvement
>  Components: Tika service connector
>Affects Versions: ManifoldCF 2.21
>Reporter: Julien Massiera
>Assignee: Julien Massiera
>Priority: Major
> Fix For: ManifoldCF next
>
>
> Some files may contain abnormally big metadata (several MB, be it for the 
> metadata values, but also for the total amount of metadata) that can be 
> problematic concerning the memory consumption of the connector. 
> To avoid this, we can provide job configuration options for the 
> TikaServiceRmetaConnector to set limits on both metadata values and global 
> amount of metadata, and exclude metadata that exceed the limits



--
This message was sent by Atlassian Jira
(v8.20.1#820001)


[jira] [Created] (CONNECTORS-1700) TikaServiceRmeta: Add options to filter out metadata based on size

2022-03-15 Thread Julien Massiera (Jira)
Julien Massiera created CONNECTORS-1700:
---

 Summary: TikaServiceRmeta: Add options to filter out metadata 
based on size
 Key: CONNECTORS-1700
 URL: https://issues.apache.org/jira/browse/CONNECTORS-1700
 Project: ManifoldCF
  Issue Type: Improvement
  Components: Tika service connector
Affects Versions: ManifoldCF 2.21
Reporter: Julien Massiera
Assignee: Julien Massiera


Some files may contain abnormally big metadata (several MB, be it for the 
metadata values, but also for the total amount of metadata) that can be 
problematic concerning the memory consumption of the connector. 

To avoid this, we can provide job configuration options for the 
TikaServiceRmetaConnector to set limits on both metadata values and global 
amount of metadata, and exclude metadata that exceed the limits



--
This message was sent by Atlassian Jira
(v8.20.1#820001)