On 4/14/20 8:17 AM, Csaba Polgar wrote:
> Hello,
> 
> In the past, there was an attempt to upgrade a Linux system from SLES11 SP4
> to SLES 12 SP4, but it was not successful because of the missing disks
> after OS upgrade.
> We saw the below error messages in the SLES11 SP4:
> 
> hostname:~ # lvs | grep "Attr\|ao--p"
>   LV                                                              VG   Attr
> LSize   Pool Origin Data%  Move Log Copy%  Convert
>   db2db_db2bmtg0_db2data                                          vg0
> -wi-ao--p  60.00g
>   db2db_db2bmtg0_db2dump                                          vg0
> -wi-ao--p   2.00g
>   is_appname                                                       vg0
> -wi-ao--p 700.00g
>   is_appname_mclm                                                  vg0
> -wi-ao--p  44.25g
>   is_appname_tss-mz                                                vg0
> -wi-ao--p 250.00g
>   opt_IBM_wstemp                                                  vg0
> -wi-ao--p  65.00g
> hostname:~ #

Actual messages from the boot process would be more helpful. This output
isn't an "error message," just a status display.

-snip-
> 2) The upgrade was very slow and struggle. (The Linux system was useable
> after 3 reboot, and there was many device(dasd)/lvm/FS error.)
> In the SLES12 I saw many simlar errror messages:
> WARNING: Device for PV x5zvM9-l42o-ifVM-zMWC-NgT5-dZ8G-afApuY not found or
> rejected by a filter.
> ...
> Couldn't find device with uuid x5zvM9-l42o-ifVM-zMWC-NgT5-dZ8G-afApuY

This is where looking at the contents of /etc/lvm/archive might be
helpful. The files in there will tell you what the device name
(/dev/dasd??) was associated with that UUID at the time it was created.
Note that this does not mean it will be the same device today. Looking
at the system logs from that time should reveal the device number
(0.0.xxxx) that device name had at the time.

> 3 ) Many error message came from the old, and unused and corrupt VG (vg0)
> and its LVs, and PVs, so I deleted these (the PVs, LVs of vg0).
>   After a new reboot, disappeared every error messages (in SLES 12 SP4,
> after upgrade).

That should have been done a long time ago.

> 4 ) At this time, I have checked every filesystem (~80, with fsck), and
> every filesystem was good.
> 
> 5 ) I updated the metadata of LVM ( to lvmetad).

I don't know what this means.

> 6 ) I performed more (~10) boot on this Linux system to test, and I saw
> similar error messages during every boot:
> [FAILED] Failed to start LVM2 PV scan on device 94:617.
> See 'systemctl status lvm2-pvscan@94:617.service' for details.

What does the output from that command show?

-snip-
> 7 ) Aftet some reboot, I saw same error messages about the new VG (vg1)
> than above (in step 2) about the old and faulty VG (vg0)!;
> It means, that one or more of the Physical Volumes is missing from the
> Linux system, from the new VG.

Do you have a list of the devices that are supposed to be there? Does
that match the list of devices that actually are there?

-snip-
> Could somebody please say me, what should be fixed to get a stable PVs,
> LVs, FSs on SLES 12?

All of this sounds like some sort of configuration problem on the
system. I've not heard of anyone else having a problem like this on
SLES12. I'm not 100% certain that's the answer, but it seems likely.

I would say you need to start with documenting the current status of the
system. What DASD devices are in use, what their device names, device
numbers, and LVM UUIDs are, etc.

Then, make backups of the system, restore them to new DASD, and document
the mapping between the two. If you're using z/VM, you should have all
the same virtual addresses as you do on the production system.

Then perform the upgrade. Make sure everything you used to have, is
still there. That means each DASD device number and the UUID associated
with it.

Investigate any messages that say something like "See 'systemctl status
lvm2-pvscan@94:617.service' for details." Examine the system log from
the boot process. Is every device being found and handled properly? And
so on. Check /etc/lvm/lvm.conf to make sure the filter statement(s) in
there aren't excluding anything you need. Essentially, you need to dig
into the whole configuration and devices to make sure everything lines
up properly and makes sense.


Mark Post

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO LINUX-390 or visit
http://www2.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to