Description of problem: The mkinitrd script uses a global variable "rootdev" to store the name of the root device. This is either auto-detected or passed in via the command line --rootdev= parameter. The changes to mkinitrd to support boot from multipath introduces code to build a list of root WWIDs (root_wwids) from a list of component root devices (rootdevs). The local variable used in the body of this loop is named "rootdev" allowing this code to clobber the global variable. This results in an incorrect root device being emitted into the /init script; instead of the root multipath partition or logical volume a component scsi device is emitted, e.g. we get: $ mkrootdev -t ext3 -o defaults,ro /dev/sdc instead of: $ mkrootdev -t ext3 -o defaults,ro /dev/mapper/mpath1p1 This prevents the system from booting. Version-Release number of selected component (if applicable): mkinitrd-5.1.19.6-44 How reproducible: Depends on system configuration; doesn't seem to affect systems with root on a logical volume. Steps to Reproduce: 1. run mkinitrd on a multipath boot system 2. unpack resulting initrd.img 3. verify root device passed on mkrootdev line Actual results: component device (/dev/sd*) Expected results: Actual root device (usually a /dev/mapper/mpath?p? device) Additional info:
I think the reason we don't always see this is here: 1250 if [ -z "$rootdevs" ]; then 1251 rootdevs="$rootdev" 1252 fi This appears right after dm device dependencies have been discovered. If we arrive here with a zero-length rootdevs then the value of rootdev at the end of the following loop is always equal to rootdev and the generated init script is correct. Otoh, if rootdevs is a list of component devices (e.g. "/dev/sdc /dev/sdf" etc.) then we get the incorrect behaviour.
Created attachment 346118 [details] rename local variable to avoid conflict with global rootdev Patch from John Ruemker to correct rootdev aliasing.
After patch review, adding dev ack for this one.
I'll take care of this one.
This is fixed in mkinitrd-5.1.19.6-55 .
~~ Attention Customers and Partners - RHEL 5.5 Beta is now available on RHN ~~ RHEL 5.5 Beta has been released! There should be a fix present in this release that addresses your request. Please test and report back results here, by March 3rd 2010 (2010-03-03) or sooner. Upon successful verification of this request, post your results and update the Verified field in Bugzilla with the appropriate value. If you encounter any issues while testing, please describe them and set this bug into NEED_INFO. If you encounter new defects or have additional patch(es) to request for inclusion, please clone this bug per each request and escalate through your support representative.
Fix confirmed in 5.5 beta. initrd-orig contains initrd contents from a 5.4 mpath install, initrd-503567 contains initrd contents created from mkinitrd-5.1.19-6.56: # grep mkrootdev initrd-orig/init initrd-503567/init initrd-orig/init:mkrootdev -t ext3 -o defaults,ro /dev/sdc initrd-503567/init:mkrootdev -t ext3 -o defaults,ro /dev/VolGroup00/LogVol00 # rpm -q mkinitrd mkinitrd-5.1.19.6-56 -John
Moving to VERIFIED as per comment #14
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHBA-2010-0295.html
*** Bug 457216 has been marked as a duplicate of this bug. ***