Release Engineering

This project is intented for all release related issues (isos, installer, etc), under the umbrella of the ArchLinux Release Engineers

FS#19454 - LVM: Entering bigger size for lv than the vg is, AIF fails

Attached to Project: Release Engineering
Opened by Lukas Grässlin (lukasg) - Friday, 14 May 2010, 10:37 GMT
Last edited by Gerardo Exequiel Pozzi (djgera) - Monday, 26 November 2012, 04:50 GMT
Task Type Bug Report
Category AIF
Status Closed
Assigned To No-one
Architecture All
Severity Medium
Priority Normal
Reported Version 2010.05.07-testbuild
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No


If I create a logical volume and the size is prompted and I enter a value that exceeds the size of the volumegroup
the installer of course fails creating the logical volume and I'm not able to proper exit or rollback the changes.
I'm in a loop: The installer always says, that it cannot create the lv and after confirming this it tries again etc.
After I flood-confirmed :) the installer exits with "ERROR: process_filesystem needs a partition as $1"
This task depends upon

Closed by  Gerardo Exequiel Pozzi (djgera)
Monday, 26 November 2012, 04:50 GMT
Reason for closing:  Deferred
Comment by Dieter Plaetinck (Dieter_be) - Wednesday, 08 December 2010, 20:34 GMT
what do you mean you can't rollback? if anything fails during filesystem creation/mounting it will ask you "do you want to rollback?" did you not get that dialog?
Comment by Dieter Plaetinck (Dieter_be) - Monday, 28 February 2011, 16:35 GMT
ah I get it. if the installer crashes, it obviously won't be there anymore to ask you whether you want to rollback.
We should:
- warn the user about big LV/VG sizes
- maybe, more properly check returncode of lvcreate/vgcreate calls, if possible.
Comment by Dieter Plaetinck (Dieter_be) - Sunday, 25 December 2011, 15:49 GMT
i just checked the logic and it seems any failure in lvcreate is properly propagate and should result in enabling you to rollback.
however, an issue we do currently have is that process_filesystems() does multiple iterations (and doesn't stop right away on failure), attempting to do the same thing a few times before finally giving up. so i'll change that to abort right away.
but still, you should see that it failed and be allowed to rollback.

i guess i'll just have to try it myself.