Bug 184412 - anaconda traceback in isValidXFS
Summary: anaconda traceback in isValidXFS
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Fedora
Classification: Fedora
Component: anaconda
Version: rawhide
Hardware: powerpc
OS: Linux
medium
medium
Target Milestone: ---
Assignee: Anaconda Maintenance Team
QA Contact: Mike McLean
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2006-03-08 16:34 UTC by Karsten Hopp
Modified: 2007-11-30 22:11 UTC (History)
1 user (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2006-03-09 16:08:47 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
anaconda traceback (39.98 KB, text/plain)
2006-03-08 16:34 UTC, Karsten Hopp
no flags Details
new traceback (52.75 KB, text/plain)
2006-03-09 11:18 UTC, Karsten Hopp
no flags Details

Description Karsten Hopp 2006-03-08 16:34:29 UTC
Description of problem:
anaconda-11.0.1-1 tracebacked during an installation on an OpenPower-720
machine. I'm using vscsi to an AIX host, the logical volumes have been newly
created.

Logfile attached...

Comment 1 Karsten Hopp 2006-03-08 16:34:30 UTC
Created attachment 125817 [details]
anaconda traceback

Comment 2 Paul Nasrat 2006-03-08 16:43:09 UTC
Manoj - can you replicate this with a box with AIX VIO here? 
Karten:

How did you create the vscsi devices?

If you switch to tty2 can you get parted /dev/sda print, if you just create a
blank partition on the vscsi drive without any volumes does it work?  Are there
any AIX error messages on the VIO server? 

Notes to self:

Is this suspect from lvm:

/dev/sda3: read failed after 0 of 2048 at 21352742912: Input/output error

16:30:02 INFO    : anaconda floppy device sda

<6>scsi0 : IBM POWER Virtual SCSI Adapter 1.5.8
<6>ibmvscsi: partner initialization complete
<6>ibmvscsi: SRP_LOGIN succeeded
<4>ibmvscsi: host srp version: 16.a, host partition IBM VIO-Server (1), OS 3,
max io 262144
<5>  Vendor: AIX       Model: VDASD             Rev:     
<5>  Type:   Direct-Access                      ANSI SCSI revision: 03
<5>SCSI device sda: 20971520 512-byte hdwr sectors (10737 MB)
<5>sda: Write Protect is off
<7>sda: Mode Sense: 17 00 00 08
<5>sda: cache data unavailable
<3>sda: assuming drive cache: write through
<5>SCSI device sda: 20971520 512-byte hdwr sectors (10737 MB)
<5>sda: Write Protect is off
<7>sda: Mode Sense: 17 00 00 08
<5>sda: cache data unavailable
<3>sda: assuming drive cache: write through
<6> sda: sda1 sda2 sda3

Comment 3 Karsten Hopp 2006-03-08 16:54:19 UTC
I've created the logical volumes in AIX with
'mklv -lv vdisk0 client_rootvg 10G'
and bound them with 'mkvdev -vdev vdisk0 -vadapter vhost0 -dev vtscsi0' to the
virtual adapter vtscsi0

client_rootvg is an AIX volumegroup with 7 physical disks

How can I switch to tty2 ? I'm running anaconda inside a vterm on the HMC and
have no idea if there are other means to do that.

Comment 4 Jeremy Katz 2006-03-08 18:22:33 UTC
Although you might hit other problems later (the read errors do generally mean
something else odd is going on), I went ahead and committed a change to catch
the read error in the fs checking code

Comment 5 Karsten Hopp 2006-03-09 11:17:09 UTC
Now I get the usual "The partition table on device sda was unreadable..... Would
you like to initialize this drine, erasing all data ?"

It tracebacked later now, not sure when as I didn't pay attention.
I'll attach the new log.

A restart of the installation after that last traceback worked fine, though. 
It looks like there is another error condition which needs to be caught by anaconda.


Comment 6 Karsten Hopp 2006-03-09 11:18:02 UTC
Created attachment 125865 [details]
new traceback

Comment 7 Jeremy Katz 2006-03-09 16:08:47 UTC
lvm is trying to read and getting read errors.  Given that it's a command line
utility with no library interface, we really can't do much other than "this failed"

But it's hitting exactly the same sort of errors reading which is either a
hardware problem or a vscsi problem (could be the AIX server or the driver on
the Linux side, but I'd lean slightly towards the server piece since it is very
fragile in my experience)


Note You need to log in before you can comment on or make changes to this bug.