[opencms-dev] Weird behavior with scheduled Lucene index building

[Mariano Barcia]

Dear Mark,

I saw this posting of yours today in the morning, and tonight we had the
opportunity to apply this procedure! The crontable got corrupted during
some indexation tests. Thank you!

One comment, maybe it contributes: this procedure is meant to be
performed while opencms is alive, isn't it? Our customer first stopped
tomcat, so after the CRONTABLE was erased, opencms refused to initialize
the servlet. However, after a little while it woke up and the crontable
was created again. So, happy end for us all, :-D

Thanks very much for this tip and documentation!
--mariano



-----Mensaje original-----
De: opencms-dev-admin at opencms.org [mailto:opencms-dev-admin at opencms.org]
En nombre de M Butcher
Enviado el: Miércoles, 28 de Enero de 2004 01:54 p.m.
Para: opencms-dev at opencms.org
Asunto: Re: [opencms-dev] Weird behavior with scheduled Lucene index
building


It is possible that you are running the CronIndexManager twice and that 
your cron table really is corrupted. I noticed last night that on 
Mozilla 1.1 the cron entries don't show up and it is possible to corrupt

the table by adding another line. Not sure why, yet, but I strongly 
suspect Javascript.

Anyway... after having to go through the trouble of manually destroying 
the cron entry, I wrote some instructions. They are in the OpenCms 
Lucene README file now, too. The crontable is stored as a serialed 
object (a blob in the db), which means you have to delete the entire 
cron table. :(

Here are the directions:

KILLING THE CRON INDEX

Every once in a while, something entered into the cron table will 
corrupt the
whole thing. As far as I can tell, there is no clean way to clear this
up.
However, if it must be done, it must be done. Here's the ugly way. YOU
WILL LOSE ALL DATA IN THE CRONTABLE... and if you screw things up, you 
could
lose more data. Just wanted to be honest about that. ;-)

0) Connect to your database.
1) select SYSTEMPROPERTY_ID, SYSTEMPROPERTY_NAME from
CMS_SYSTEMPROPERTIES;

You should see one named CRONTABLE. That holds a serialized Java class
that
contains the contents of the evil corrupt crontable. It must be deleted.

2) delete from CMS_SYSTEMPROPERTIES where SYSTEMPROPERTY_NAME =
'CRONTABLE';
3) You don't even have to restart -- a new CRONTABLE entry will be 
generated
by the OpenCms Scheduled Tasks manager.




Roth Olivier wrote:
> Hi,
> 
> I have a strange problem with the daily rebuilding of the Lucene
index.
> I have one entry in scheduled tasks:
> 05 01 * * * admin Administrators
> net.grcomputing.opencms.search.lucene.CronIndexManager
createIndex=true
> 
> Sometimes it works, sometimes it doesn't and I get an
> ArrayIndexOutOfBoundException.
> When I look closer at the log, I see that every file is being indexed
twice,
> as if 2 instances of the CronIndexManager where running
simultaneously.
> Now I get a lots of messages saying "crontable is corrupt. Scheduler
is now
> disabled!", even after a restart of tomcat.
> Where can I delete these entries ?
> 
> Greetings,
> 	Olivier
> 
> _______________________________________________
> This mail is send to you from the opencms-dev mailing list
> To change your list options, or to unsubscribe from the list, please
visit
> http://mail.opencms.org/mailman/listinfo/opencms-dev

_______________________________________________
This mail is send to you from the opencms-dev mailing list
To change your list options, or to unsubscribe from the list, please
visit
http://mail.opencms.org/mailman/listinfo/opencms-dev