Bug 21045

Summary:	Could not mount compressed loopback
Product:	Mageia	Reporter:	Pierre Fortin <pf>
Component:	Installer	Assignee:	Mageia tools maintainers <mageiatools>
Status:	RESOLVED FIXED	QA Contact:
Severity:	normal
Priority:	Normal	CC:	marja11
Version:	Cauldron	Keywords:	NEEDINFO
Target Milestone:	---
Hardware:	All
OS:	Linux
Whiteboard:
Source RPM:		CVE:
Status comment:

Description Pierre Fortin 2017-06-08 00:51:37 CEST

Description of problem:  Trying to install fresh mga6 (cauldron) from up-to-date local copy of mirrors. For the first time ever, I now get "Could not mount compressed loopback", followed by "Could not load program into memory"


Version-Release number of selected component (if applicable): 


How reproducible:  First saw this when trying to install onto new SSD drive today. Reproduced trying to install into VBox.  


Steps to Reproduce:
1. Start with boot.iso
2. Select repository
3. Image is loaded, then gives error "Could not mount compressed loopback :(."

Comment 1 Marja Van Waes 2017-06-08 23:18:12 CEST

I'm not sure I understand how far you get.... does "Image is loaded" mean that stage2 was started?

If so, can you switch to tty2, attach a USB key and type

      bug

and check whether report.bug was written to the USB key?

If that doesn't work, but you do see log files in /tmp/ then please attach those.

CC: (none) => marja11
Keywords: (none) => NEEDINFO
Assignee: bugsquad => mageiatools

Comment 2 Pierre Fortin 2017-06-09 14:07:28 CEST

I've done numerous installs on many machines over the past 19 years and never saw this error before; not even when it was reported circa 2006-2009.

>1. Start with boot.iso
>2. Select repository
    Installation method:  ->Hard disk
    Select disk:  ->sdb
    Select partition:  ->sdb7
    Directory:  distros/6.64
>3. Image is loaded, "Loading program into memory..." (with red progress bar)
>          then gives error "Could not mount compressed loopback :(."

grep'ed the install directory for "compressed" and "loopback"; but could not find the above error string. Since I'm using VBox to reproduce, I can provide screen shots if desired...

tty2
bug: command not found 

tty3: 
[...]
* assuming media/distros/6.64 is a mirror tree     # it is
* Total Memory: 134217720 MBytes
* mount_compressed_may_preload: /tmp/images/install/stage2//mdkinst.sqfs into /tmp/stage2 (preload - 1)
* /dev/loop0 is available
* mounting /dev/loop0 on /tmp/stage2 as type squashfs
* mount failed: Invalid argument
* unsetting automatic

Comment 3 Pierre Fortin 2017-06-11 13:09:11 CEST

SIGH...  I seem to be able to find the most obscure bugs...

The source of this issue was the local repo disk partition was full and the daily update failed to complete, leaving the compressed loopback in an incomplete/corrupted state.  Cleared some space, getting the local repo in a proper state and it's working now...

Rather than "Could not mount compressed loopback :(."; a better message might include "Missing, corrupted, incomplete due to disk full, ..."

Resolution: (none) => FIXED
Status: NEW => RESOLVED