[ome-users] restarting ome server(s) - mac os x

[Chris Allan]

On Thu, May 26, 2005 at 11:52:51AM -0700, Bob Marinelli wrote:
> Hi Ilya & Chris,

Hi again Bob.

> 
> The httpd/error_log is pretty clean, for each image being imported we  
> get this:
> 

... snip.

> but we do have a fair number of postgres tasks in unix top:

... snip.

A good idea if this happens is to load up psql and run:

ome=# SELECT * FROM pg_stat_activity;

... snip.

> Upon browsing over to http://localhost, we get this (have not seen it  
> 
> Error serving OME::Web::Home: DBD::Pg::db rollback failed: rollback  
> failed at /Library/Perl/5.8.6/OME/Factory.pm line 574.
> Internal Error

If you restart Postgres, since OME caches database handles, you must
restart Apache. OME can't rollback these transactions because the
database handles it held are now invalid.

> 
> And the httpd log had an exceptionally interesting set of errors. A  
> restart was the next
> step, this cleared up things nicely. It was then possible to log in  
> and clear the Tasks items,
> the hung tasks showed up as Failed and it was possible to clean them.
> 
> >As Chris said, we haven't done testing on Tiger, but are glad to  
> >see its not completely broken.
> >I am curious to see what the last few hundred lines of /var/log/ 
> >httpd/error_log has to say though.
> >-Ilya
> 
> Thanks to your very quick responses, I'll keep a log of anything we  
> come across that looks like
> a tiger-specific problem.
> 
> For today, I'll try tracking down the import failures in ImageManager.

Good luck. :)

> 
> Again, thanks very much,
> 
> Bob
> 

Ciao.

-Chris