Re: nosh: service-dt-scanner gets repeatedly killed by SIGABRT

From: Guillermo <gdiazhartusch_at_gmail.com>
Date: Wed, 8 Jul 2015 22:00:23 -0300

Sorry about that, I accidentally sent this before finishing. So again:

2015-07-08 5:30 GMT-03:00 Jonathan de Boyne Pollard:
>
> If there's a SIGABRT there should be some error output saying why.

Nope, no error output. I have lauched service-dt-scanner from a shell
in the foreground, and to be sure, I also had service-manager (in turn
launched from a shell) supervise the service-dt-scanner process and
report, using this restart script:

#!/bin/execlineb -S0
foreground { echo service-dt-scanner: $1 $2 $3 }
exit 1

So both service-manager and service-dt-scanner had their stdout and
stderr connected to the terminal, and I should have been able to see
any error messages. I did see indeed service-manager's "INFO:" and
"DEBUG:" messages. But no messages about why service-dt-scanner died.
The shell just said "Aborted", "echo $?" returned 134, and the restart
script above said "service-dt-scanner: abort ABRT 6".

2015-07-08 5:30 GMT-03:00 Jonathan de Boyne Pollard:
>
> This:
>>
>> Some of the actions that make service-dt-scanner die are using the
>> service-{control,status,show,is-*} commands on the service, and using ls on
>> its bundle directory (yeah, listing its contents).
>
> is completely mad. Another process merely listing the contents of the
> directory shouldn't be an event visible to service-dt-scanner in the first
> place. My immediate suspicion is a (fourth) libkqueue bug.

I will add that using ls on the scan directory makes
service-dt-scanner output this message: "service-dt-scanner: DEBUG:
event filter 0 ident 6 fflags 6". So libkqueue does likely generate an
event when ls is used on a watched directory.

I'll post later what strace had to say.

Thank you,
G.
Received on Thu Jul 09 2015 - 01:00:23 UTC