[Bug 1121720] New: Retaining existing Software RAID via AutoYast is broken: y2storage always formats MD
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Bug ID: 1121720
Summary: Retaining existing Software RAID via AutoYast is
broken: y2storage always formats MD
Classification: openSUSE
Product: openSUSE Tumbleweed
Version: Current
Hardware: x86-64
OS: Other
Status: NEW
Severity: Major
Priority: P5 - None
Component: YaST2
Assignee: yast2-maintainers@suse.de
Reporter: pieter@hollants.com
QA Contact: jsrain@suse.com
Found By: ---
Blocker: ---
Created attachment 794273
--> http://bugzilla.suse.com/attachment.cgi?id=794273&action=edit
Autoyast profile to reproduce; install once first, then try to re-install
Retaining an existing software RAID via AutoYast is also broken with regard to
reuse of the MD itself: inspite of reformat == false being picked up from the
Autoyast profile, y2storage md_creator.rb wants to format an MD device because
of this code block:
if planned_md.partitions.empty?
format_md(new_graph, md, planned_md)
else
partition_md(new_graph, md, planned_md)
end
...because planned_md.partitions is empty. I couldn't figure out yet whether
planned_md SHOULD have contained the member partitions or whether the
if-statement needs modification.
NOTE: The attached example Autoyast file uses
"
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c1
Steffen Winterfeldt
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c2
Steffen Winterfeldt
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Steffen Winterfeldt
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c3
Ancor Gonzalez Sosa
Ancor, probably the same as bug 1120979?
Yes, I'm pretty sure this is a duplicate of bug#1120979. (In reply to Pieter Hollants from comment #0)
NOTE: The attached example Autoyast file uses "
/dev/md/0 " for both the "/dev/md" drive as well as the RAID member partitions to work around the bug in #1120979.
The style of the names used in the profile (/dev/md/0 vs /dev/md0) is not
relevant, since (as already mentioned in the other bug report) the real problem
doesn't come from matching the strings in the profile with the device name.
The problem is how we translate this to a device name in the (Auto)YaST code:
<drive>
<device>/dev/md</device>
<partitions>
<partition>
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c4
--- Comment #4 from Pieter Hollants
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Arvin Schnell
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c5
--- Comment #5 from Ancor Gonzalez Sosa
I'm wondering where the confusion could come from in the first place: so far I've seen the kernel only use "/dev/md0", so I assume "/dev/md/0" is some YaST-internal syntax?
No, it's not. According to libstorage-ng, that's the real kernel name of the MD device. Ask Arvin for more details. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c6
--- Comment #6 from Arvin Schnell
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c7
--- Comment #7 from Pieter Hollants
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Ancor Gonzalez Sosa
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c9
--- Comment #9 from Ancor Gonzalez Sosa
I don't think it is a duplicate: by looking at
Yes, this is just another consequence of the very same bug. The same fix that we already reported for bug#1120979 (adding a slash to the line linked below) also fixes this other symptom. https://github.com/yast/yast-storage-ng/blob/master/src/lib/y2storage/autoin...
I deducted that specifying an explicit
in the section, I could work around the other bug #1120979:
Wrong deduction ;-) Reusing a raid was broken either by using /dev/md0 and by using /dev/md/0. With the fix we already commented in the other bug, both work.
<drive> [...]
/dev/md/0 [...] </drive>But then I discovered _this_ bug.
Then you rediscovered the same bug. ;-)
Please also note that by referring to the raid as /dev/md/0 instead of /dev/md0 I think we've broken compatibility with existing Autoyast profiles: both the Autoyast Documentation for openSUSE 15.0 at https://doc.opensuse.org/projects/autoyast/#ay.partition_sw_raid and the newest documentation I could find, for SLES 12SP4, use "/dev/md0" as
.
As said, backwards compatibility is not broken. Once the bug is fixed, it will work for both /dev/md0 and /dev/md/0. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c10
Ancor Gonzalez Sosa
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c12
Ancor Gonzalez Sosa
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Swamp Workflow Management
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Swamp Workflow Management
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Swamp Workflow Management
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Swamp Workflow Management
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Swamp Workflow Management
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Swamp Workflow Management
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c17
--- Comment #17 from Swamp Workflow Management
http://bugzilla.suse.com/show_bug.cgi?id=1121720
http://bugzilla.suse.com/show_bug.cgi?id=1121720#c18
--- Comment #18 from Swamp Workflow Management
http://bugzilla.suse.com/show_bug.cgi?id=1121720
Swamp Workflow Management
participants (1)
-
bugzilla_noreply@novell.com