Bug 699164

Summary: MDRaidError: mddeactivate failed for /dev/md3: mdadm: failed to stop array /dev/md3: Device or resource busy Perhaps a running process, mounted filesystem or active volume group?
Product: [Fedora] Fedora Reporter: Bevis King <brwk>
Component: anacondaAssignee: David Lehman <dlehman>
Status: CLOSED WONTFIX QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: medium Docs Contact:
Priority: unspecified    
Version: rawhideCC: anaconda-maint-list, jonathan, vanmeeuwen+fedora
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard: anaconda_trace_hash:e7f7698360239fb1354fcc8f9828ac73bfec24ed7ae14dd44213e4187b6ed7dd
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2011-05-03 15:36:56 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
Attached traceback automatically from anaconda. none

Description Bevis King 2011-04-23 20:38:40 UTC
The following was filed automatically by anaconda:
anaconda 14.22 exception report
Traceback (most recent call first):
  File "/usr/lib64/python2.7/site-packages/pyanaconda/storage/devicelibs/mdraid.py", line 202, in mddeactivate
    raise MDRaidError("mddeactivate failed for %s: %s" % (device, msg))
  File "/usr/lib64/python2.7/site-packages/pyanaconda/storage/devices.py", line 2876, in teardown
    mdraid.mddeactivate(self.path)
  File "/usr/lib64/python2.7/site-packages/pyanaconda/storage/__init__.py", line 1301, in findExistingRootDevices
    device.teardown()
  File "/usr/lib64/python2.7/site-packages/pyanaconda/upgrade.py", line 110, in findExistingRoots
    return findExistingRootDevices(anaconda, upgradeany=upgradeany)
  File "/usr/lib64/python2.7/site-packages/pyanaconda/upgrade.py", line 81, in findRootParts
    flags.cmdline.has_key("upgradeany"))
  File "/usr/lib64/python2.7/site-packages/pyanaconda/dispatch.py", line 212, in moveStep
    rc = stepFunc(self.anaconda)
  File "/usr/lib64/python2.7/site-packages/pyanaconda/dispatch.py", line 131, in gotoNext
    self.moveStep()
  File "/usr/lib64/python2.7/site-packages/pyanaconda/gui.py", line 1278, in setScreen
    self.anaconda.dispatch.gotoNext()
  File "/usr/lib64/python2.7/site-packages/pyanaconda/gui.py", line 1191, in nextClicked
    self.setScreen ()
MDRaidError: mddeactivate failed for /dev/md3: mdadm: failed to stop array /dev/md3: Device or resource busy
Perhaps a running process, mounted filesystem or active volume group?

Comment 1 Bevis King 2011-04-23 20:38:49 UTC
Created attachment 494456 [details]
Attached traceback automatically from anaconda.

Comment 2 Brian Lane 2011-04-25 18:40:56 UTC
Can you provide some more detail about the system? It looks like you were trying to upgrade. There are also a number of ldconfig errors in the logs, but I am not sure if they are related.

Comment 3 Bevis King 2011-05-03 13:04:27 UTC
OK.  The system is based upon a Dual-Core 64-bit Intel Atom chipset and has five 250GB SATA notebook hard drives with a Promise TX SATA PCI controller in the one PCI slot in the motherboard.  Each drive was identically partitioned with a small slice at the beginning and "the rest".  On drives sda1 and sdb1 the root/boot partition was mirrored as /dev/md3.  On drives sdd1 and sde1 there is a RAID0 making a larger /tmp as /dev/md1.  Swap is directly on sdc1.  The remaining large partition on each drive was combined into a RAID5 array as /dev/md2 which was a single large LVM2 volume.

I tried to upgrade from Fedora 12 to Fedora 14.  Anaconda really didn't understand the disc layout at all and bombed out in many different ways.  The various bug reports I filed were the various failures as I used the ALT-F2 console to re-mount the various partitions/slices in the hope that it would aid anaconda in performing the upgrade.

Ultimately all my efforts were failures and I installed a fresh Fedora 14 onto a new manually created /dev/md0 (on /dev/sda1 and /dev/sdb1) and then progressively relocated the other partitions onto the LVM area and other md devices.

Probably the best fix for anaconda might be some way in which you could let it proceed with the package update efforts without messing with the disc layout.  It works, it's valid, but it is pretty complex.

Regards, Bevis.

Comment 4 Bevis King 2011-05-03 13:05:28 UTC
I think the ldconfig errors are all related to some zero-size files in the Oracle 11g installation on this system.  They've been there for a while and persist after the fresh install of Fedora 14.

Regards, Bevis.

Comment 5 David Lehman 2011-05-03 14:57:04 UTC
So you went to the shell and manually activated your md and lvm and then returned to the installer? This is not supported. If you want to attach logs from a run in which you let the installer discover the devices without your intervention I can take a look at them to see what may be going on.

Are you aware that Fedora 15 will be out soon? A beta is already available.

Comment 6 Bevis King 2011-05-03 15:08:54 UTC
I'm pretty sure one of the earliest bugs submitted is exactly that scenario.  Basically I tried the simple method first and anaconda blew up big time.

Comment 7 David Lehman 2011-05-03 15:36:56 UTC
Ok, we're going to close the others and work on the one that's a supported attempt.