Hi,

On Thu, 2005-03-31 at 19:32 +0200, Arno Lehmann wrote:
> Brice Figureau wrote:
> > I'm happily running Bacula 1.36.1 (switching to 1.36.2 asap), but I have
> > the following problem.
> > My bacula director backups a few windows user workstations every
> > evening.
> > Some of the employees of the company forget to leave their workstation
> > up and running before leaving in the evening, and thus shutdown them.
> > 
> > Why is the director spending those large amount of time to retry this
> > host ?
> 
> Basically, there arereasons for it. You will find some answers in the 
> list archive, and you can change the directors configuration.
> 
>  From the manual for 1.36.0:
> 
> FD Connect Timeout = <time>
>      where time is the time that the Director should continue attempting 
> to contact the File daemon to start a job, and after which the Director 
> will cancel the job. The default is 30 minutes.
> 
> > Why is he waiting that amount of time between each access ?
> > 
> > And finally how can I control:
> >  1) the amount of time between unsuccessfull attempts
> 
> Reschedule On Error = <yes/no>
[snip]
> Reschedule Times = <count>
>      This directive specifies the maximum number of times to reschedule 
> the job. If it is set to zero (the default) the job will be rescheduled 
> an indefinite number of times.
> Priority = <number>
> 
> >  2) the number of unsuccessfull attempts before giving a fatal error 
> 
> Also helpful in a scenario like yours:
> Rerun Failed Levels = <yes/no>
>      If this directive is set to yes (default no), and Bacula detects 
> that a previous job at a higher level (i.e. Full or Differential) has 
> failed, the current job level will be upgraded to the higher level. This 
> is particularly useful for Laptops where they may often be unreachable, 
> and if a prior Full save has failed, you wish the very next backup to be 
> a Full save rather than whatever level it is started as.

This is really helpfull. I was not aware of this setting (and I thought
I read the manual thoroughly...)

[snip]
> To completely disable timeouts, you can also use a Run Before Script 
> which tries to ping the clients and immediately aborts with an exit code 
>  >0 and probably some useful output. That's what I use.

Thank you for your insightfull answers.

Before I got your answer I implemented the Run Before Script feature
with the help of the checkhost script which can be found in Bacula's
examples directory (except that most of those damn windows workstations
runs the WinXP firewall wich was forbiding icmp echo reply ).

That's also too bad that the 'Run Before Script' can't access to the
Client hostname or address. I should have planned more carefully my
Client's name so that they match the DNS of the client...

I'm now activating the 'Rerun Failed Levels'.

Take care,
-- 
Brice Figureau
Days of Wonder



-------------------------------------------------------
This SF.net email is sponsored by Demarc:
A global provider of Threat Management Solutions.
Download our HomeAdmin security software for free today!
http://www.demarc.com/Info/Sentarus/hamr30
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Reply via email to