Tomcat hangs deleting big sites from share

[braguzz]

configuration:
4.2.c community
Ubuntu server 12.04 64 8Gb 8 CPU vsphere virtual


problem:
deleting a site with many nodes share tell me 'unable to delete site' , tomcat hang and cisf and http connection stop to work.

in catalina.log I see

—
2013-07-02 08:22:05,161 INFO [webscripts.connector.RemoteClient] [http-apr-80-exec-12] Error status 408 Read timed out
java.net.SocketTimeoutException: Read timed out
at java.net.SocketInputStream.socketRead0(Native Method)
at java.net.SocketInputStream.read(SocketInputStream.java:150)
at java.net.SocketInputStream.read(SocketInputStream.java:121)
.
.
.
2013-07-02 08:22:05,192 ERROR [extensions.webscripts.AbstractRuntime] [http-apr-80-exec-12] Exception from executeScript - redirecting to status template error: 06020001 Failed to execute script 'classpath*:alfresco/site-webscripts/org/alfresco/modules/delete-site.post.json.js': 06020000 TypeError: Cannot read property "success" from null (file:/media/Data01/alfresco-4.2.c/tomcat/webapps/share/WEB-INF/classes/alfresco/site-webscripts/org/alfresco/modules/delete-site.post.json.js#12)
org.springframework.extensions.webscripts.WebScriptException: 06020001 Failed to execute script 'classpath*:alfresco/site-webscripts/org/alfresco/modules/delete-site.post.json.js': 06020000 TypeError: Cannot read property "success" from null (file:/media/Data01/alfresco-4.2.c/tomcat/webapps/share/WEB-INF/classes/alfresco/site-webscripts/org/alfresco/modules/delete-site.post.json.js#12)
ecc..
—

I've tried to install it with alfresco installer and with manual install (starting from a clean machine following <a>http://francescocorti.wordpress.com/2013/01/09/installation-alfresco-4-2-c-on-ubuntu/</a> ), but the problem persists.

I've already tried to change tomcat connectionTimeout and JAVA_OPTS without luck.

[scouil]

Hi,

As a workaround, you could try removing nodes bits by bits from the document library and only use the delete site button when the site isn't that big anymore.

[braguzz]

tomcant hangs also when someone delete from share a folder with many nodes.

It's impossible for me to put in production the server because if someone delete those 'big' folder, the other have no more access to the whole documents…

I've notice that when server 'hangs' I see many postgress process with cpu 100%

thanks in advance,
PB

[scouil]

Well yes, if you want big capacity you need to put big means behind.
You know, you can have alfresco, share, solr, postgres all on different servers.

And if you want even more, you can use clustering (only available on enterprise version I think). And use load balancers to redirect traffic between your nodes.

Alfresco scales well, but you have to give it the possibility to do so.

[braguzz]

thanks Scouil, when I say 'big' I say 4Gb, and I suppose that a linux with 8 core with 8Gb should not crash.

and then it seems that there is some misconfiguration (i hope), because I dont think that Alfresco allow that an user could crash the system simply erasing a folder…

[scouil]

You said that postgresql was the one freezing. So it's not about the size of the file I think.
It's more the number of files/folders .
For each one of them he'll check all the permissions to see if you can actually delete it based on your user/groups.
But it might also be some kind of conflict evolving into a deadlock. Try monitoring postgres and see what's actually running.

[braguzz]

I think that pistgress CPU goes high after tomcat timeout.

It sound me strange that i'm tue only one with this problem ….

[braguzz]

Do you think the problem disappares if. I install another server for solr index and postgresql? Otherwise whhat can. I do?

Thanks in advance,  Paolo

[braguzz]

Hi,
I made some investigation in postgresql log, and it seems that start to erasing but when timeout occours it start to make rollback
and cpus goes 100%.

[braguzz]

ERROR: duplicate key value violates unique constraint "idx_alf_conturl_cr"
DETAIL: Key (content_url_short, content_url_crc)=(2eedaafc.bin, 692580845) already exists.
STATEMENT: insert into alf_content_url (id, content_url, content_url_short, content_url_crc, content_size, orphan_time)
values ($1, $2, $3, $4, $5, $6)It was postgres.

Installing MySQL problem disappeared.

Thanks, paolo.
Tomcat hangs becouse of an endless list of: