On Tue, 1 Mar 2016 21:29:44 +0000 Jonathan de Boyne Pollard wrote:

> Francesco Poli:
> 
> > Please help me understand, in case I am completely off-track.
> >
> 
> You are off-track.  This is not an mdadm bug.  There is nothing to fix 
> in mdadm for this.

Mmmmh, I am not too convinced...

As I said, during an upgrade of the mdadm package, "invoke-rc.d mdadm
stop" and "invoke-rc.d mdadm start" are executed. These two commands
are completely ineffective, if systemd is PID 1. The old version of the
mdadm monitor process continues to run, unless the superuser manually
issues "service mdmonitor restart" after the upgrade.

On the other hand, if sysvinit is PID 1, the mdadm monitor process is
automatically restarted during an mdadm package upgrade.

Hence, I think that there is indeed something to be fixed in the mdadm
Debian package: perhaps, the mdadm.service symbolic link to /dev/null
could be removed and mdmonitor.service could be renamed as
mdadm.service ...

> This is not a systemd bug.  There is nothing to fix 
> in systemd for this, either.

I never claimed that this was a bug in systemd: I filed this bug report
against the mdadm package.

> 
> Francesco Poli:
> 
> > Unfortunately, checkrestart does not suggest it.
> >
> 
> This is not a checkrestart bug even.  There is nothing to fix in 
> checkrestart for this ...
> 
> > https://anonscm.debian.org/cgit/collab-maint/debian-goodies.git/commit/?id=aed5839eb8511eade953b82dd80dd8c855cc4f42
> >
> 
>   ... because it is fixed already, over a year ago.  You are using 
> debian-goodies version 0.63 or earlier.  Upgrade to debian-goodies 0.64 
> or later.

Nope, wrong guess, sorry.
I am using debian-goodies version 0.64:

  $ dpkg -l | grep debian-goodies | cut -c 1-60
  ii  debian-goodies                          0.64

and I have always used this version of debian-goodies on the box where
I experienced the mdadm misbehavior:

  $ zgrep debian-goodies /var/log/aptitude*
  /var/log/aptitude.5.gz:[INSTALL] debian-goodies:amd64 0.64



-- 
 http://www.inventati.org/frx/
 There's not a second to spare! To the laboratory!
..................................................... Francesco Poli .
 GnuPG key fpr == CA01 1147 9CD2 EFDF FB82  3925 3E1C 27E1 1F69 BFFE

Attachment: pgp8bvZbQdi_4.pgp
Description: PGP signature

Reply via email to