Delete Client fails

Bug #565166 reported by smallming
12
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Openbravo ERP
In Progress
Undecided
Gorka Gil

Bug Description

Delete Client fails after waiting for about half a minute at the processing screen and redirects back to login page.

Upon login as System Administrator, the client still remains in the system.

A new client was created during Initial Client Setup, without Accounting File.
System is ubuntu 10.04 beta 2, setup according to instructions on http://planet.openbravo.com.

Revision history for this message
Juan Pablo Aroztegi (jpabloae) wrote :

Hi. What error message do you get? Could you take a screenshot? Also, could you attach the Tomcat logs? They're locate at /opt/OpenbravoERP-2.50/logs/openbravo.log and /opt/OpenbravoERP-2.50/logs/catalina.out.

visibility: private → public
Changed in openbravo-erp:
assignee: nobody → Juan Pablo Aroztegi (jpabloae)
Revision history for this message
smallming (smallming) wrote :
Revision history for this message
smallming (smallming) wrote :
Revision history for this message
smallming (smallming) wrote :
Revision history for this message
smallming (smallming) wrote :

I've reproduced the error just before saving the logs so the relevant log entries may be found at the bottom.

Revision history for this message
dr.spock (dr.spock) wrote :

I have had the same problem under Ubuntu Server 10.04 64 bits. I switched back the Openbravo user to English (USA) and then it shows an error instead going back to the login page:

Error:
out of shared memory

I have googled a little and it seems it's something related to the PostgreSQL memory management (maybe the default configuration can't handle so many tables or data as Openbravo needs). I didn't install PG by myself, it was installed as a dependency of the openbravo-erp package.

As I'm not very familiarized with PostgreSQL, could someone tell me where and how sould I increase that "shared memory"?

Thanks in advance.

Revision history for this message
dr.spock (dr.spock) wrote :

Solved.

I added the following to the /etc/sysctl.conf:

************
kernel.shmmax=134217728
kernel.shmall=134217728
************

And I modified /etc/postgresql/8.4/openbravo-erp/postgresql.conf from

************
shared_buffers = 24MB # min 128kB
************

to be:

************
shared_buffers = 64MB # min 128kB
************

It failed again first time I tried, but next two times went well. I don't know if those values aren't still enough or they're oversized, it would be nice to know from OB developers.

Changed in openbravo-erp:
assignee: Juan Pablo Aroztegi (jpabloae) → Gorka Gil (gorka.gil)
status: New → Confirmed
status: Confirmed → In Progress
Revision history for this message
Juan Pablo Aroztegi (jpabloae) wrote :

dr. spock,

Sorry about the delay in the response and thank you for the information you are providing. This is indeed the root issue: having 2 PostgreSQL clusters started at the same time (main and openbravo-erp), the system does not have enough max shared memory enabled. In principle the way we'll fix it in the package is as you propose, but creating a /etc/sysctl.d/openbravo-erp.conf instead of modifying /etc/sysctl.conf.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.