PDA

View Full Version : sles 11 sp1 to sp2 upgrade failure using patchCD or upgrade



ramesh_tn
09-Mar-2012, 16:10
I am just testing the upgrade to SLES 11 SP2 from SP1 using a virtual image of a running physical machine. But I keep encountering boot corruption. Whichever method I choose, either UPGRADE after booting of SP2 media or PATCHCD in Yast using the same SP2 media, results in the system booting to a GRUB prompt and halting. While the install or upgrade is in progress, I can see some messages saying "mkinitrd failed ". I cant try this on the actual server till this issue is resolved. My config is as follows:

Root partition is non-lvm partition using Ext3. All the remaining partitions use LVM. Directories USR, OPT, SRV, VAR etc are all separately mounted LVM partitions on HP Raid array.

Another thing I noticed, is that during the upgrade procedure after booting of the SP2 media, I only see the USR and VAR partitions being used. The rest are not even mounted. I checked this by logging into a new terminal and using "df -h", though the purpose was to check the free space being used up during upgrade.

Please advise, what the problem could be.

System is a old HP Proliant ML 350 g4.

ramesh_tn
10-Mar-2012, 13:01
Correction in above post: Please note the physical system specifications (raid, hardware model) are really not relevant as this a virtualbox guest. I inadvertently used the original physical machine specs. The virtual machine uses LSI scsi controller. The rest remain.

KBOYLE
10-Mar-2012, 21:09
ramesh tn wrote:

> >
> > Correction in above post: Please note the physical system
> > specifications (raid, hardware model) are really not relevant as
> > this a virtualbox guest. I inadvertently used the original physical
> > machine specs. The virtual machine uses LSI scsi controller. The
> > rest remain.

You posted in the suse.support.server.install-boot forum. Are you
trying to do this on a SLES box? If so, virtualbox MAY work but it is
not supported. Please tell us what your host operating system is.

I don't have any experience with virtualbox and don't even know if
SLES11-SP2 is even supported when run in virtualbox. Do you?

>
> I am just testing the upgrade to SLES 11 SP2 from SP1 using a virtual
> image of a running physical machine. But I keep encountering boot
> corruption.

Are You sure your virtual image is good? Have you done this before?

> Whichever method I choose, either UPGRADE after booting of
> SP2 media or PATCHCD in Yast using the same SP2 media, results in the
> system booting to a GRUB prompt and halting.

At this point, the only supported way to upgrade is to boot from the
SLES11-SP2 media. Online update should be available within the next
week or weeks.


> My config is as follows:
>
> Root partition is non-lvm partition using Ext3. All the remaining
> partitions use LVM. Directories USR, OPT, SRV, VAR etc are all
> separately mounted LVM partitions on HP Raid array.
>
> Another thing I noticed, is that during the upgrade procedure after
> booting of the SP2 media, I only see the USR and VAR partitions being
> used. The rest are not even mounted. I checked this by logging into a
> new terminal and using "df -h", though the purpose was to check the
> free space being used up during upgrade.
>
> Please advise, what the problem could be.
>
> System is a old HP Proliant ML 350 g4.

Even with the "correction" you provided, and which I included at the
top of this post, your configuration is not clear.

When you partition a RAID array, those partitions are seen by the host
OS as separate disk drives. They are sometimes called Logical Volumes
or LUN's. SLES also has a Logical Volume Manager (LVM) that allows you
to create Logical Volumes within a partition. These LV's have some of
the same characteristics as a partition: for example, they can be used
for USR, OPT, SRV, VAR etc. mount points.

If you can describe your virtual SLES11-SP1 configuration a bit better,
someone may be able to offer you some suggestions.

--
Kevin Boyle - Knowledge Partner
If you find this post helpful and are using the web interface,
show your appreciation and click on the star below...

ramesh_tn
12-Mar-2012, 12:42
I have provided as much info as I can next to your replies.

Thank you for taking the time to help out. My formatting effort has unfortunately mixed up my reply within yours. Apologize for that. Little unsure about how to use the formating options here.


ramesh tn wrote:

> >
> > Correction in above post: Please note the physical system
> > specifications (raid, hardware model) are really not relevant as
> > this a virtualbox guest. I inadvertently used the original physical
> > machine specs. The virtual machine uses LSI scsi controller. The
> > rest remain.

You posted in the suse.support.server.install-boot forum. Are you
trying to do this on a SLES box? If so, virtualbox MAY work but it is
not supported. Please tell us what your host operating system is.

Host OS is Win 7 64 bit. Guest is SLES 11 SP1 64 bit


I don't have any experience with virtualbox and don't even know if
SLES11-SP2 is even supported when run in virtualbox. Do you?


Dont know about SP2 but have been running SP1 without issues.


>
> I am just testing the upgrade to SLES 11 SP2 from SP1 using a virtual
> image of a running physical machine. But I keep encountering boot
> corruption.

Are You sure your virtual image is good? Have you done this before?

> Whichever method I choose, either UPGRADE after booting of
> SP2 media or PATCHCD in Yast using the same SP2 media, results in the
> system booting to a GRUB prompt and halting.

At this point, the only supported way to upgrade is to boot from the
SLES11-SP2 media. Online update should be available within the next
week or weeks.

I tried the boot method albeit within Virtualbox. Does not work. Whats more it corrupts boot record.


> My config is as follows:
>
> Root partition is non-lvm partition using Ext3. All the remaining
> partitions use LVM. Directories USR, OPT, SRV, VAR etc are all
> separately mounted LVM partitions on HP Raid array.
>
> Another thing I noticed, is that during the upgrade procedure after
> booting of the SP2 media, I only see the USR and VAR partitions being
> used. The rest are not even mounted. I checked this by logging into a
> new terminal and using "df -h", though the purpose was to check the
> free space being used up during upgrade.
>
> Please advise, what the problem could be.
>
> System is a old HP Proliant ML 350 g4.

Even with the "correction" you provided, and which I included at the
top of this post, your configuration is not clear.

When you partition a RAID array, those partitions are seen by the host
OS as separate disk drives. They are sometimes called Logical Volumes
or LUN's. SLES also has a Logical Volume Manager (LVM) that allows you
to create Logical Volumes within a partition. These LV's have some of
the same characteristics as a partition: for example, they can be used
for USR, OPT, SRV, VAR etc. mount points.

If you can describe your virtual SLES11-SP1 configuration a bit better,
someone may be able to offer you some suggestions.

Below is the output of df -h which shows all partitions and lvms clearly.
Filesystem Size Used Avail Use% Mounted on
/dev/sda1 2.0G 378M 1.6G 20% /
devtmpfs 3.9G 280K 3.9G 1% /dev
tmpfs 3.9G 940K 3.9G 1% /dev/shm
/dev/mapper/system-home1 485M 11M 449M 3% /home
/dev/mapper/system-local1 97M 5.6M 87M 7% /local
/dev/mapper/system-opt1 245M 6.1M 226M 3% /opt
/dev/mapper/system-srv1 245M 11M 222M 5% /srv
/dev/mapper/system-tmp1 1008M 37M 921M 4% /tmp
/dev/mapper/system-usr1 5.0G 4.1G 673M 86% /usr
/dev/mapper/system-var1 2.0G 847M 1.1G 45% /var
/dev/mapper/system-vmslvg1 30G 3.7G 25G 14% /vms
/dev/mapper/system-smtrepo1 20G 13G 6.4G 66% /srv/www/htdocs/repo


--
Kevin Boyle - Knowledge Partner
If you find this post helpful and are using the web interface,
show your appreciation and click on the star below...

KBOYLE
12-Mar-2012, 18:51
ramesh tn wrote:

> Host OS is Win 7 64 bit. Guest is SLES 11 SP1 64
> > bit
> >
> >
> > I don't have any experience with virtualbox and don't even know if
> > SLES11-SP2 is even supported when run in virtualbox. Do you?
> >
>
> Dont know about SP2 but have been running SP1 without
> > issues.
> >

SLES 11 SP2 was released, based on kernel version 3.0.10. It would be
interesting to know if a new install of SLES11-SP2 will run on
virtualbox before trying to diagnose issues when updating from
SLES11-SP1.

Other than that, I'm afraid I'm out of ideas.

--
Kevin Boyle - Knowledge Partner
If you find this post helpful and are using the web interface,
show your appreciation and click on the star below...

Magic31
12-Mar-2012, 21:46
Another thing I noticed, is that during the upgrade procedure after booting of the SP2 media, I only see the USR and VAR partitions being used. The rest are not even mounted. I checked this by logging into a new terminal and using "df -h", though the purpose was to check the free space being used up during upgrade.


As Kevin already states, doing this through VirtualBox might not be the most representable way to do this.

In any case: if when booting from the SLES 11 SP2 DVD to upgrade, and you are not seeing all partitions mounted properly... that would cause problems. The SRV mount point would be a lesser one of importance... all the others are needed for the upgrade to succeed.

Is the upgrade screen showing the correct root partition to be upgraded? Or do you have to manually choose it by first showing all partitions?

-Willem

ramesh_tn
14-Mar-2012, 08:00
During install no errors are displayed. The root partition is detected and so all are the LVMs. What I noticed was only some of them were mounted during the upgrade procedure. I noticed the "mkinitrd" failure in the small area that displays the package update progress. Should'nt something like this produce a more siginificant "STOP" error??

Will also check the suggestion in the previous post and see if a direct install works on Virtualbox.

KBOYLE
03-Apr-2012, 05:18
ramesh tn wrote:

> Whichever method I choose, either UPGRADE after booting of
> SP2 media or PATCHCD in Yast using the same SP2 media, results in the
> system booting to a GRUB prompt and halting.

When your system halts, you don't by chance see a message saying
"invalid root filesystem", do you? If so, that is a known problem that
Suse is working to resolve. It has nothing to do with running in a VM.

If this is your issue, you can try to type "exit" then <enter> and the
boot process may continue.

--
Kevin Boyle - Knowledge Partner
If you find this post helpful and are using the web interface,
show your appreciation and click on the star below...