Re: Timeout waiting for /dev/md/imsm0 ?

[Mariusz Tkaczyk <mariusz.tkaczyk@linux.intel.com>]

On Tue, 16 Aug 2022 19:04:02 -0700
"David F." <df7729@gmail.com> wrote:

> What rules should be used?   I don't see a /dev/md directory, I
> created one, stopped the raid (all the /dev/md* devices went away)
> and tried to start the raid, same thing and only /dev/md127 gets
> created, nothing in /dev/md/ directory and none of the md126 devices ?
>  You then get the timeout.

https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/tree/udev-md-raid-arrays.rules#n24
ENV{DEVTYPE}=="disk", ENV{MD_DEVNAME}=="?*", SYMLINK+="md/$env{MD_DEVNAME}"

The clue here is:
IMPORT{program}="BINDIR/mdadm --detail --no-devices --export $devnode"
where property MD_DEVNAME is generated. It is obtained from "/var/run/mdadm/map"
file and generally is same as name property in metadata.

It is all in udev-md-raid-arrays.rules

Mariusz
> 
> On Tue, Aug 16, 2022 at 12:03 AM Mariusz Tkaczyk
> <mariusz.tkaczyk@linux.intel.com> wrote:
> >
> > Hi David,
> > On Mon, 15 Aug 2022 15:28:08 -0700
> > "David F." <df7729@gmail.com> wrote:
> >  
> > > I'm not sure if this list is getting the messages but to summarize, if
> > > I pass the 5.15.x kernel parameter "nomdraid" to skip udev from doing
> > > anything with the RAID and then run:
> > >
> > > mdadm - v4.1 - 2018-10-01
> > >
> > > mdadm --examine --scan to create the /etc/mdadm/mdadm.conf file with:
> > >
> > > ARRAY metadata=imsm UUID=788c3635:2e37de4b:87d08323:987f57e5
> > > ARRAY /dev/md/TestRAID container=788c3635:2e37de4b:87d08323:987f57e5
> > > member=0 UUID=835de710:3d35bfb1:d159af46:6570f120
> > >
> > > Then run:
> > >
> > > mdadm --assemble --scan --no-degraded -v
> > >
> > > I get:
> > >
> > > mdadm: looking for devices for further assembly
> > > mdadm: no RAID superblock on /dev/sdc1
> > > mdadm: /dev/sdc is not attached to Intel(R) RAID controller.
> > > mdadm: No OROM/EFI properties for /dev/sdc
> > > mdadm: no RAID superblock on /dev/sdc
> > > mdadm: no RAID superblock on /dev/sda5
> > > mdadm: no RAID superblock on /dev/sda3
> > > mdadm: no RAID superblock on /dev/sda2
> > > mdadm: no RAID superblock on /dev/sda1
> > > mdadm: /dev/sdb is identified as a member of /dev/md/imsm0, slot 1.
> > > mdadm: /dev/sda is identified as a member of /dev/md/imsm0, slot 0.
> > > mdadm: added /dev/sdb to /dev/md/imsm0 as 1
> > > mdadm: added /dev/sda to /dev/md/imsm0 as 0
> > > mdadm: Container /dev/md/imsm0 has been assembled with 2 drives
> > > mdadm: timeout waiting for /dev/md/imsm0
> > > mdadm: looking for devices for /dev/md/TestRAID
> > > mdadm: cannot open device /dev/md/imsm0: No such file or directory
> > > mdadm: Cannot assemble mbr metadata on /dev/sdc1
> > > mdadm: Cannot assemble mbr metadata on /dev/sdc
> > > mdadm: /dev/sdb has wrong uuid.
> > > mdadm: Cannot assemble mbr metadata on /dev/sda5
> > > mdadm: Cannot assemble mbr metadata on /dev/sda3
> > > mdadm: Cannot assemble mbr metadata on /dev/sda2
> > > mdadm: no recogniseable superblock on /dev/sda1
> > > mdadm: /dev/sda has wrong uuid.
> > > mdadm: looking for devices for /dev/md/TestRAID
> > > mdadm: cannot open device /dev/md/imsm0: No such file or directory
> > > mdadm: Cannot assemble mbr metadata on /dev/sdc1
> > > mdadm: Cannot assemble mbr metadata on /dev/sdc
> > > mdadm: /dev/sdb has wrong uuid.
> > > mdadm: Cannot assemble mbr metadata on /dev/sda5
> > > mdadm: Cannot assemble mbr metadata on /dev/sda3
> > > mdadm: Cannot assemble mbr metadata on /dev/sda2
> > > mdadm: no recogniseable superblock on /dev/sda1
> > > mdadm: /dev/sda has wrong uuid.
> > >
> > > If I let UDEV start it and then stop the RAID with:
> > >
> > > mdadm --stop --scan  
> >
> > No, You didn't ask udev. You asked mdadm to do clean up. It will trigger
> > "remove" event at some point so then udev will be involved.  
> > >
> > > (which does stop it) then try to start again using the above command,
> > > I still get the timeout.
> > >
> > > This was working fine with older version 5.10.x kernel with the
> > > following differences:
> > >
> > >    mdadm v4.1 - 2018-10-01 (but from a different build - debian
> > > instead of devuan)
> > >    kmod as an older version
> > >    udev (eudev) was built against the older kmod.
> > >    all the various shared libraries and utilities were moved up to
> > > versions with Devuan Chimaera
> > >    rules updated (although I tried with the old rules too, no difference)
> > >
> > > Any idea on what is wrong?  Any tricks to have it output more
> > > information to diagnose what is happening?   The /dev/md127 device
> > > gets created, the actual devices never get created, even if you wait.  
> >
> > The error you mentioned in subject is caused by udev. This error means that
> > udev failed to create link in /dev/md/ directory.
> > If you are not referencing to this link, it can be ignored. We are expecting
> > that udev will create the link and we are waiting for it as some point in
> > mdadm.
> >
> > Thanks,
> > Mariusz