On Tue, Nov 21, 2017 at 12:57:00 -0500, Nathan Stratton Treadway wrote:
> > I do not see how Amanda::Util::setup_application call code in holding.c
>
> To be precise what I actually saw was activity involving the "pid" files
> in the holding directory showing up in an strace log. I am not sure how
> it got called, but assume it was coming from holding.c because it had
> the arguments for the kill syscall in the correct order (and thus I
> believe the activity couldn't have been from the then-in-place
> Holding.pm code). I will investigate further and let you know what I
> find.
>
Okay, sorry for the confusion -- now that I have paid careful attention
to the part of amflush prior to the Holding::get_all_datestamps() call,
I see that it's not setup_application() that updates the pid files, but
rather Disklist:add_holding_to_disklist().
>From there the program flow appears to be:
Amanda::Logfile::search_holding_disk -> find.c:search_holding_disk()
-> holding.c:holding_get_files() -> holding.c:holding_walk() ->
holding_dir_stop_if_pid_fn() -> take_holding_pid()
(Obviously the above flow only applies after the holding-pid-3.5.diff
patch has been applied.)
In any case, the pid file is updated by this call in the early part of
the amflush run (see below), which explains why the patch for the
Holding.pm logic needed to include support for the pid file pointing to
current/parent pid....
(In order to confirm the origin of the pid updates, I added
system("find /amanda/TestBackup-holding -name pid -ls");
and a print-message line immediately before and after the call to
add_holding_to_disklist(), and here is the resulting output:
=====
# su backup -lc "amflush TestBackup"
5636101 4 -rw------- 1 backup backup 5 Nov 21 18:29
/amanda/TestBackup-holding/20171121145009/pid
5636099 4 -rw------- 1 backup backup 5 Nov 21 18:29
/amanda/TestBackup-holding/20171120172656/pid
**** calling Disklist:add_holding_to_disklist ****
**** returned from Disklist:add_holding_to_disklist ****
5636101 4 -rw------- 1 backup backup 5 Nov 21 18:32
/amanda/TestBackup-holding/20171121145009/pid
5636099 4 -rw------- 1 backup backup 5 Nov 21 18:32
/amanda/TestBackup-holding/20171120172656/pid
Flushing dumps from 20171121145009 using storage "TestBackup", tapechanger "".
To volume TESTBACKUP-06 or a new volume, (The last dumps were to volume
TESTBACKUP-03
Flushing dumps from 20171121145009 using storage "TestOffsite", tape changer
"doxpop_offsite_vtapes".
To volume TESTBACKUP-109 or a new volume, (The last dumps were to volume
TESTBACKUP-03
Are you sure you want to do this [yN]?
=====
[I know from my earlier debugging that the value put into the updated
pid file is actually the parent pid.])
Nathan
----------------------------------------------------------------------------
Nathan Stratton Treadway - [email protected] - Mid-Atlantic region
Ray Ontko & Co. - Software consulting services - http://www.ontko.com/
GPG Key: http://www.ontko.com/~nathanst/gpg_key.txt ID: 1023D/ECFB6239
Key fingerprint = 6AD8 485E 20B9 5C71 231C 0C32 15F3 ADCD ECFB 6239
