Red Hat Bugzilla – Bug 470726
aic7xxx/AHA-2940U2W not catch up initialize disks, following nash's mkblkdevs is ineffective
Last modified: 2008-12-15 04:17:09 EST
Description of problem:
i686 system with AHA-2940U2W controller, 2x MAXTOR ATLAS10K4 wide-SCSI disks in SW RAID1 mode. Installation OK (release Nov-3-2008, kernel 22.214.171.124-68.fc10.i686, mkinitrd-6.0.69-1.fc10.i386), but system not booted, as system disks are not
on time detected in initrd. Update to 126.96.36.199-79.fc10.i686 and mkinitrd-6.0.70-1.fc10.i386 has no effect.
It seems aic7xxx driver needs some delay to successful disks init.
When I insert into initrd's init script 5 sec sleep (3 secs was too little)
between "modprobe aic7xxx" and subsequent "mkblkdevs", then all things
passed OK. I'm not sure about other, better solution (In man pages undocumented "mkblkdevs" nash command should be responsible for it?). My fix:
echo Creating block device nodes.
echo "Loading raid1 module"
modprobe -q raid1
echo "Loading raid456 module"
modprobe -q raid456
echo "Loading scsi_transport_spi module"
modprobe -q scsi_transport_spi
echo "Loading aic7xxx module"
modprobe -q aic7xxx
sleep 5 +++++++++++ Added delay
mdadm -As --auto=yes --run /dev/md0
Version-Release number of selected component (if applicable):
Steps to Reproduce:
There are a number of bugs like this, in Fedora, and several filed under RHEL too.
You can add "MODULES=scsi_scan_wait" to /etc/sysconfig/mkinitrd. This will load the module designed for this waiting, and you won't have to edit init by hand.
For (much) more background, see
(In reply to comment #1)
> You can add "MODULES=scsi_scan_wait" to /etc/sysconfig/mkinitrd. This will
Sorry, that should be MODULES=scsi_wait_scan
Yes, it's right solution. But if it was here some /etc/sysconfig/mkinitrd
template, life would be little lighter.
This bug appears to have been reported against 'rawhide' during the Fedora 10 development cycle.
Changing version to '10'.
More information and reason for this action is here:
This bug has already been reported, so I'm closing this as a dup of the earlier report. Note that I believe we have a fix for this available now, see the comments I'm about to add to bug 466607.
*** This bug has been marked as a duplicate of bug 466607 ***