On Thu, Jan 25, 2018 at 04:16:52PM -0600, Eric Sandeen wrote:

On 1/25/18 3:44 PM, Darrick J. Wong wrote:

quoted

quoted

Ok this seems batshit crazy, no offense.  I don't blame /you/ ;)  You pointed me
at  http://refspecs.linuxbase.org/LSB_2.0.1/LSB-PDA/LSB-PDA/iniscrptact.html
but it says that the /initscript/ should exit with specific codes, not that
the application it /calls/ should do so.

Yes.  At the moment, the systemd service calls xfs_scrub directly, hence
it interprets the return code as an initscript error code.  So in theory
we could create a wrapper script that does all the is_service junk, but
now that's another weird little script to break.  On the plus side we'd
contain the systemd workaround crap to some random wrapper in /usr/lib.

Ok, I'll think about it.

does LSB even have anything to do w/ systemd behavior or have they reinvented
all that by now?  :(

I think systemd is mostly following LSB here, though they give us the
flexibility to screw things up.

FWIW,

https://www.freedesktop.org/software/systemd/man/systemd.service.html

says that you can define:

SuccessExitStatus=

    Takes a list of exit status definitions that, when returned by the main service process, will be considered successful termination

does that help?

No.  The xfs_scrub_fail@ service is only invoked if the service fails,
so if we use SuccessExitStatus= to remap the scrub problem error codes
to "success" then systemd considers the service to have succeeded and
won't start the _fail service.

--D

-Eric
--
To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html