Red Hat Bugzilla – Bug 503681
kickstart loses existing LVM config
Last modified: 2010-11-28 10:44:48 EST
Created attachment 346197 [details]
ks to reuse existing lvm vg
Description of problem:
Kickstart installations from the F11-rc2 DVD which attempt to reuse existing LVM volume groups (volgroup --useexisting) will fail with an error that the VG name is not recognized. Anaconda cannot reuse LVM PV partitions in any existing volume groups. Semi-automatic kickstart installations with manual layout also exhibit the issue. This bug did not exist with the F11-preview DVD.
Version-Release number of selected component (if applicable):
220.127.116.11 (i386 RC2 DVD)
Steps to Reproduce:
1. create a ks.cfg which reuses existing volume groups
2. install with the kickstart
The following error was found while parsing your kickstart configuration
The following problem occurred on line 17 of the kickstart file
No preexisting VG with the name "vg0" was found
The same kickstart works fine with F11-preview.
This is yet another regression from Preview to RC2.
Before F11 installation, the disk is laid out as follows:
sda3 lvm pv of volume group vg0
Volume group vg0 contains
f9 - the F9 /
free space to become the F11 /
Created attachment 346198 [details]
i386 Preview anaconda.log (ks works)
Created attachment 346199 [details]
i386 Preview program.log (ks works)
Created attachment 346200 [details]
i386 Preview storage.log (ks works)
Created attachment 346201 [details]
i386 Preview syslog (ks works)
Created attachment 346202 [details]
i386 RC2 anaconda.log (ks doesn't work)
Created attachment 346203 [details]
i386 RC2 program.log (ks doesn't work)
Created attachment 346204 [details]
i386 RC2 storage.log (ks doesn't work)
Created attachment 346205 [details]
i386 RC2 syslog (ks doesn't work)
I looked into it and it seems we are not setting DeviceTree.clearPartType correctly somewhere (it is set to default CLEARPART_TYPE_LINUX, while it should be CLEARPART_TYPE_NONE), so the existing vg device is not added to devicetree during storage initialization because it it doesn't pass this check:(storage.devicetree.py:1529)
if shouldClear(device, self.clearPartType, self.clearPartDisks,
# if this is a partition that will be cleared by clearpart,
# don't bother with format-specific processing
The clearPartType setting became hard to follow in the code through the storage rewrite, and as Chris did some changes of it recently, I think he'll know better where/how to change it (if I am right).
I posted a patch for review. It makes attached ks work for me. You can try this updates file for version .58 of anaconda containing the patch:
Re: Comment 10
Confirmed. The patch works for me. I actually tried it with two ks files, the one included and the one I'll actually use when I install real systems.
You can set this bug to MODIFIED, too.
Thanks for the effort.
Will this fix be on the GA media (if you know)? If so, I'll update the wiki page with that info now and remove to the bug from the wiki once I confirm the media have (or at least the i386 medium has) the fix.
Thanks again. Now on to try out the RAID bug...
Should be fixed in next build of anaconda for rawhide.
Unfortunately it is too late for F11.
This bug appears to have been reported against 'rawhide' during the Fedora 11 development cycle.
Changing version to '11'.
More information and reason for this action is here:
*** Bug 504913 has been marked as a duplicate of this bug. ***
Re: Comment 36
Moving from F11 back to rawhide. It only makes sense there.
Re: Comment 15
Oops. Cut & Paste. I meant Re: Comment 13, not Re: Comment 36 (which doesn't exist yet here).
*** Bug 505575 has been marked as a duplicate of this bug. ***
According to koji, there have been no builds for anaconda since 2 June.
Therefore this bug cannot be fixed yet.
Am I missing something?
*** Bug 508455 has been marked as a duplicate of this bug. ***
I do experience the same problem.
Does the bug fix ?
(In reply to comment #20)
> I do experience the same problem.
> Does the bug fix ?
For F11, there's not a genuine bug fix. You have a few options. I would recommend, in the order listed:
1. Don't use kickstart.
2. Try the anaconda update from Comment 10, although it's older than the
version of anaconda on the GA media, so there's no guarantee it doesn't
regress other bug fixes that made it on to the GA media. If the update is
still available by the time anyone reads this comment, this option is the
closest thing to a bug fix for F11 until ...
3. Wait for a respin, which will probably have an updated anaconda.
For Rawhide, I'm downloading today's boot.iso now. Today's boot.iso/install.img should include yesterday's anaconda build (12.1), so I'm hopeful.
Unable to verify fix in anaconda-12.0 due to Bug 509572
Unable to verify fix in anaconda-12.1 due to Bug 510172
Putting to modified, fix has been pushed into anaconda 12.0.
I was able to verify the fix in anaconda-12.3 with today's (21 July 2009) boot.iso.
It looks like I hit this bug in Fedora 14 again. Here is my partitioning from the kickstart config file:
# Use all existing partition schema and reformat all but /home
partition /boot --usepart=sda1
partition pv.6fZc4L-gojR-fTgc-Twhq-LcaN-XYc8-yeomBJ --usepart=sda2
volgroup vg_prhost2 pv.6fZc4L-gojR-fTgc-Twhq-LcaN-XYc8-yeomBJ --useexisting
logvol / --vgname=vg_prhost2 --name=lv_root --useexisting
logvol /home --vgname=vg_prhost2 --name=lv_home --useexisting --noformat
logvol swap --vgname=vg_prhost2 --name=lv_swap --useexisting
[root@prhost2 ag]# pvdisplay
--- Physical volume ---
PV Name /dev/sda2
VG Name vg_prhost2
PV Size 465.27 GiB / not usable 23.00 MiB
Allocatable yes (but full)
PE Size 32.00 MiB
Total PE 14888
Free PE 0
Allocated PE 14888
PV UUID Jlexvs-KWlQ-dhR9-vdNr-yUrK-QbAg-2wjcPg
[root@prhost2 ag]# vgdisplay
--- Volume group ---
VG Name vg_prhost2
Metadata Areas 1
Metadata Sequence No 4
VG Access read/write
VG Status resizable
MAX LV 0
Cur LV 3
Open LV 3
Max PV 0
Cur PV 1
Act PV 1
VG Size 465.25 GiB
PE Size 32.00 MiB
Total PE 14888
Alloc PE / Size 14888 / 465.25 GiB
Free PE / Size 0 / 0
VG UUID UU4f01-PIEA-VvSL-7bxb-Mm3k-DVVc-gnQ97J