[OpenIndiana-discuss] what pkg data can safely be removed?
Tim Mooney
Tim.Mooney at ndsu.edu
Tue Jun 17 02:47:03 UTC 2025
In regard to: Re: [OpenIndiana-discuss] what pkg data can safely be...:
> In message <924b035f-bda5-6ae5-ef0e-d05a08de7aec at ndsu.edu>, Tim Mooney via open
> indiana-discuss writes:
>> $ sudo find /var/pkg/publisher -print | wc -l
>> 1465282
>
> I don't recall seeing it documented, but I have, so far without
> noticeable ill-effect, occasionally done:
> # pkg unset-publisher openindiana.org
> # rm -rf /var/pkg/publisher/openindiana.org
> # pkg set-publisher -g http://pkg.openindiana.org openindiana.org
That's a great trick John, and I would have used it, but TIL that
setting 'flush-content-cache-on-success' cleans out /var/pkg/publisher/,
*not* /var/pkg/cache.
When I did a pkg uninstall after setting
pkg set-property flush-content-cache-on-success True
pkg "hung" for more than 20 minutes at the end:
Actions: 1/127 actions (Removing old actions)
Actions: Completed 127 actions in 0.14 seconds.
Done (0.604s)
Done (0.001s)
Done (1.475s)
Done (55.556s)
Done (0.185s)
Done (0.000s)
<--- "hung" here
Done (1278.296s)
I figured out while waiting for the last "Done" that it was cleaning out
/var/pkg/publisher. That directory looks much better now:
$ sudo du -hs /var/pkg/publisher
676M /var/pkg/publisher
$ sudo find /var/pkg/publisher -print | wc -l
13953
> Not seen it documented and haven't had the bandwidth to explore
> the code, but my WAG is /var/pkg/publisher/ is a cache publisher's
> package files to speed the creation of zones.
Based upon the name of the setting that cleans it out, seems your guess
is correct.
I'm still not sure about the ancient stuff in the /var/pkg/license
directory, but at least the /var/pkg/publisher directory appears to be
getting some maintenance now.
Tim
--
Tim Mooney Tim.Mooney at ndsu.edu
Enterprise Computing & Infrastructure /
Division of Information Technology / 701-231-1076 (Voice)
North Dakota State University, Fargo, ND 58105-5164
More information about the openindiana-discuss
mailing list