Big news! The community will be moving to a new platform April 21. Read more.
Big news! The community will be moving to a new platform April 21. Read more.
Absent Member.
Absent Member.
1393 views

preparing for oes 11

In the preparation documents it says that linux mount points need to be done by id or path. On my xen vm guest that is an oes2 sp 2 version, I attempted to change to id and I could see the letter designation change from K (for name) to I or P (for id or path). But when I apply and finish and go back in I see the K for name is back. It seems the changes are not sticking.

This is using yast2 disk from the terminal to launch the gui to make the changes. Suggestions.
Labels (2)
0 Likes
6 Replies
Absent Member.
Absent Member.

Users of virtual environments such as XEN or VMware should use either mount by UUID or mount by LABEL as the /dev/disk/by-id/ devices contain hardware information from the physical disk (e.g. serial number) which is not available in virtual environments

TID 7003734
0 Likes
Knowledge Partner Knowledge Partner
Knowledge Partner

jmcg wrote:

>
> In the preparation documents it says that linux mount points need to
> be done by id or path. On my xen vm guest that is an oes2 sp 2
> version, I attempted to change to id and I could see the letter
> designation change from K (for name) to I or P (for id or path).
> But when I apply and finish and go back in I see the K for name is
> back. It seems the changes are not sticking.
>
> This is using yast2 disk from the terminal to launch the gui to make
> the changes. Suggestions.


It's even worse than that!

When running in a DomU, /dev/disk may only contain by-uuid (by-id and
by-path may not be present). If by-id is not present and customer
chooses by-id mount option, or if by-path is not present and customer
chooses by-path mount option, partitioner will change fstab and set the
device field to blanks for the specified mount point.

[Example]
Before change:

/dev/xvda3 / reiserfs acl,user_xattr 1 1
/dev/xvda1 /boot ext3 acl,user_xattr 1 2
/dev/xvda2 swap swap defaults 0 0

After change:

/ reiserfs acl,user_xattr 1 1
/dev/xvda1 /boot ext3 acl,user_xattr 1 2
/dev/xvda2 swap swap defaults 0 0
[/Example]

This will cause a boot failure.

Once the device name in fstab has been set to blanks, partitioner is
unable to repair fstab but instead exhibits the following behaviour:

1. Partitioner cannot change the mount option back to Device Name
because it thinks the device is already mounted by Device Name.

2. Partitioner will allow the mount option to be changed to UUID
however it doesn't change the existing invalid device in fstab.
Instead, it creates an additional entry in fstab at the same mount
point as the invalid device.

This behaviour can be seen in SLE10, SLE11, and SLE11-SP1 and possibly
SLE11-SP2. It is documented in Bug #761226.

Both the SLES and OES documentation still currently insist that mount
point options be changed to by-id or by-path and say nothing about
alternatives for virtual machines.

TID 7003734 states "Users of virtual environments such as XEN or VMware
should use either mount by UUID or mount by LABEL" but it does not
include the additional information I have provided above.


--
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...
_____
Kevin Boyle - Knowledge Partner - Calgary, Alberta, Canada
Who are the Knowledge Partners?
If you appreciate my comments, please click the Like button.
If I have resolved your issue, please click the Accept as Solution button.
0 Likes
Knowledge Partner Knowledge Partner
Knowledge Partner

I know in XEN (at least on SLES 10 XEN Dom0), I've always done /dev/disk/by-id (I'm using raw disk, not file backed)

BUT, in the DomU, I'll just let is use the /dev/xvda

why?

Because the disks on the Dom0 are already defined by id:
/dev/disk/by-id/bighairy number = xvda

But that's probably not the case for everyone (not everyone wants to use RAW disks/LUN's). I cannot speak to vmware as I've never put anything on it that didn't have a single solitary disk.

What's odd, however, is that if the SLES 11 SP1 OS install "needs" to have /dev/disk/by-id, then why on earth would it (SLES) not write that information into the fstab?
0 Likes
Knowledge Partner Knowledge Partner
Knowledge Partner

kjhurni wrote:

> What's odd, however, is that if the SLES 11 SP1 OS install "needs" to
> have /dev/disk/by-id, then why on earth would it (SLES) not write that
> information into the fstab?


A DomU doesn't need to have "/dev/disk/by-id". The issue is that if you
ever needed to add a second disk to the DomU, say for more space, that
new disk could be presented to the DomU as xvda and the original would
become xvdb and the DomU probably wouldn't boot.

Using any mount option, other than Device Name, would prevent that from
happening.

--
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...
_____
Kevin Boyle - Knowledge Partner - Calgary, Alberta, Canada
Who are the Knowledge Partners?
If you appreciate my comments, please click the Like button.
If I have resolved your issue, please click the Accept as Solution button.
0 Likes
Knowledge Partner Knowledge Partner
Knowledge Partner

KBOYLE;2196685 wrote:
kjhurni wrote:

> What's odd, however, is that if the SLES 11 SP1 OS install "needs" to
> have /dev/disk/by-id, then why on earth would it (SLES) not write that
> information into the fstab?


A DomU doesn't need to have "/dev/disk/by-id". The issue is that if you
ever needed to add a second disk to the DomU, say for more space, that
new disk could be presented to the DomU as xvda and the original would
become xvdb and the DomU probably wouldn't boot.

Using any mount option, other than Device Name, would prevent that from
happening.

--
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...


hmmm, interesting. Never had that happen, but then again I've only ever added a new disk in XEN via the host and it got an explicit /dev/disk/by-id that then went to /dev/xvdb

Of course, most of my XEN domU are still OES2 SP3, so I've not yet had the joy of OES11/SLES11 yet.

Thanks for the TID info, BTW. Now I'm even more confused (haha)

So do we submit TID feedback to get the TID corrected or something?
(an idiot like me would blindly follow said TID And possibly run into the bug you mentioned and then I'd not be very happy)
0 Likes
Knowledge Partner Knowledge Partner
Knowledge Partner

kjhurni wrote:

> So do we submit TID feedback to get the TID corrected or something?


Let's see what happens to Bug #761226. I'm working with NTS on a couple
of things and hopefully everything will be sorted out.

--
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...
_____
Kevin Boyle - Knowledge Partner - Calgary, Alberta, Canada
Who are the Knowledge Partners?
If you appreciate my comments, please click the Like button.
If I have resolved your issue, please click the Accept as Solution button.
0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.