Bug#782793: systemd: ext4 filesystem on lvm on raid causes boot to enter emergency shell

[Rick Thomas]

On Apr 23, 2015, at 2:31 PM, Michael Biebl <biebl at debian.org> wrote:

> Am 23.04.2015 um 21:47 schrieb Rick Thomas:
>> 
>> On Apr 23, 2015, at 1:29 AM, Michael Biebl <biebl at debian.org> wrote:
>> 
>>> There might be a race somewhere, i.e. lvm2-activation(-early).service
>>> being run *before* mdadm has assembled the RAID.
>>> 
>>> You could test this theory, be artifically delaying those two services.
>>> Copy them to /etc/systemd/system, and add a ExecStartPre=/bin/sleep 30
>>> to the [Service] section of those two units.
>> 
>> 
>> So… That worked (adding a 30-second delay) twice in a row.
>> 
>> Let me know if you want to see a log with debug turned on for such a reboot.
>> 
> 
> Ok, I guess I have bad news then.
> lvm2 in jessie is not really hotplug aware. The
> lvm2-activation(-early).service units run at a static, fixed time during
> boot. You'd need something like lvm2's lvmetad support [1]. I tried to
> enable that in Debian today, but failed miserably due to various bugs
> (already filed 3 bugs today regarding lvmetad).
> Ultimately, this is something which needs to be fixed in lvm2 and at
> this point I don't think there is a lot more that I can do from the
> systemd side.
> 
> Sorry for the unsatisfying answer.
> 
> Regards,
> Michael
> 
> 
> 
> [1]
> https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html/Logical_Volume_Manager_Administration/metadatadaemon.html

No reason to be sorry!

First of all, you have given me two solid work-arounds to the problem (#1 the 30-second delay you suggested; #2 the cron reboot script I wrote based on your preliminary suggestions.)

Second, you have filed bug reports that (hopefully) will be a prod to get the real problem worked on by those folks who are able to actually fix it.

This is what the open-source philosophy is all about: Given enough eyeballs, all bugs are shallow!  We each contributed a pair of eyeballs to this bug.  Now it’s time for others to contribute theirs.

Early on, I cloned this bug as 782865 and re-assigned it to lvm2.  You may want to close the clone and re-assign this one to lvm2 so the folks in that part of the woods have access to the full record.  I’ll leave that up to you because you probably know more about the proper procedures and protocols than I do.

Thanks *very* much for all your patient help!

Enjoy!
Rick