[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#921741: cups: Cups sheduler stops with program error when using JobPreserveHistory <seconds>



On Mon 04 Mar 2019 at 19:39:24 +0100, Emin Kaya wrote:

> Hi Brian,

Hello Emin. Thank you for the level of detail.
 
> Am Mo., 4. März 2019 um 16:49 Uhr schrieb Brian Potkin <
> claremont102@gmail.com>:
> 
> > [...]
> 
> 
> 
> > > To do this, we found "JobPreserveHistory <seconds>" as root option in
> > > the documentation for /etc/cups/cupsd.conf, which would do the job
> > > perfectly.
> >
> > PreserveJobHistory?
> >
> 
> Yes, you're right. I shouldn't send out mail after a long shift.. Upps.

Just checking. It seemed like a typo. Try not to work so hard.

> [...]
> 
> > > within the last 86400 seconds, the /var/log/cups/error.log contain the
> > > Error message
> > > "E:[Timpestamp] cups sheduler stops through a program error"
> >
> > I cannot reproduce this. Do you still experience it?
> >
> 
> We still have this behaviour, we tested it either with "1d" for one day and
> also with
> the equivalent "86400" in seconds.
> 
> We tried it via the webinterface, as also by manually modifying
> /etc/cups/cupsd.conf
> and restarting cups with "systemctl restart cups".
> 
> In both cases, the job history is empty and the error_log contain this
> error message.
> 
> [...]
> 
> > From my tests it seems that PreserveJobHistory cannot handle being
> > given any seconds value, small or large. What do you get with 30 for
> > PreserveJobHistory?
> >
> 
> I will test it, but if it work that way, either the implementation is buggy
> or the documentation
> for this option isn't correct... I prefer the first one to be fixed. ;-)
> 
> Edit: I tested "PreserveJobHistory" with several values: 60, 1024, 86400 by
> manually editing /etc/cups/cupsd.conf
> and after each modification, I did a service restart.

I used 'cupsctl PreserveJobHistory=30', but it amounts to the same thing.

> Expecting: JobHistory will listing test print jobs within the specified
> time frame, not listing them after the given time frame exeeds
> 
> What happened: In each case, all job files (*.c, *.d) under /var/spool/cups
> are deleted, even the creation date of the file was not exeeding the tested
> value.
> Instead of originally reported, the error_log contained no error after a
> service restart.
> 
> I did not test also by changing the values via the webinterface.
> 
> PreserveJobHistory seems not handling numeric values as otherwise described
> as possible value in the cupsd.conf(5)-documentation.
> Instead, it handle numeric values like the option value "No" and disable
> the job history completely.

That is more or less the conclusion I came to too. I will forward this
report upstream tomorrow after doing a few more tests.

-- 
Brian.


Reply to: