Re: shellcommands not executing - lock problem?

[Russell Adams]

I was thinking that perhaps there was a limit, like the lockname was
truncated before comparison...

Looking in lock.c (no expert, but thought I'd check the code before I
opened my mouth), it often uses the first 100 chars of the operator,
so you'd get:

shellcommand._gweb._bin_su___gweb__l__c___usr_bin_rsync__rltvv___delete_gweb_nue_gentoo_org__var_www_www_gentoo_org_stage_site_images___var_www_www_gentoo_org_images__

truncated to:

shellcommand._gweb._bin_su___gweb__l__c___usr_bin_rsync__rltvv___delete_gweb_nue_gentoo_org__var_www_www_gentoo_org_stag

which would mean the locks were most likely identical for those two
operations.

The path munging workaround should still work.

Am I totally offbase above?

Russell

On Tue, Feb 24, 2004 at 06:21:14PM +0100, Mark.Burgess@iu.hio.no wrote:
> 
> Ah - perhaps it is because the strings are so similar, and cfagent restricts 
> the
> keys length....they end up with the same key..
> Try reordering the options on one of the lines...?
> 
> M
> 
> On 24 Feb, Kurt Lieber wrote:
> > On Tue, Feb 24, 2004 at 05:51:22PM +0100 or thereabouts, 
> > Mark.Burgess@iu.hio.no wrote:
> >> If something caused the agent to crash while the shellcommand was running
> >> the lock will have to wait until expiredafter time is reached before
> >> it will be removed. You can always delete the lock database to check.
> >> /var/cfengine/cfengine_lock_db
> > 
> > That's the odd thing -- the second command *never* gets run while the other
> > two *always* get run, even though they're in the same shellcommands stanza.
> > We've been trying to make it work for a couple weeks now.
> > 
> >> > shellcommands:
> >> >   _gweb.slave_webnode.slave_web_sync::
> >> >     "/bin/su - gweb -l -c '/usr/bin/rsync -rltvv --delete 
> >> > gweb@$(_master_node):$(web_stage_html)/ $(web_html)/'"
> >> >     "/bin/su - gweb -l -c '/usr/bin/rsync -rltvv --delete 
> >> > gweb@$(_master_node):$(web_stage_images)/ $(web_images)/'"
> >> >     "/bin/su - gweb -l -c '$(web_scripts)/rsync-weblogs.sh'"
> > [snip]
> >> > cfengine:nue: Nothing scheduled for 
> >> > shellcommand._gweb._bin_su___gweb__l__c___
> >> > usr_bin_rsync__rltvv___delete_gweb_nue_gentoo_org__var_www_www_gentoo_org_stage
> >> > _site_images___var_www_www_gentoo_org_images__ (0/1 minutes elapsed)
> > 
> > Just to make sure, I deleted /var/cfengine/cfengine_lock_db on one of the
> > servers and re-ran the command.  Same problem.
> > 
> > Also, this behavior is consistent across all three boxes that this
> > particular stanza runs on.  It *never* works on any of the boxes unless we
> > use the -K option.
> > 
> > --kurt
> 
> 
> 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Work: +47 22453272            Email:  Mark.Burgess@iu.hio.no
> Fax : +47 22453205            WWW  :  http://www.iu.hio.no/~mark
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> 
> 
> _______________________________________________
> Help-cfengine mailing list
> Help-cfengine@gnu.org
> http://mail.gnu.org/mailman/listinfo/help-cfengine