Bug 615998
| Summary: | Kickstart installation on IMSM RAID volume fails if an old metadata was present on drives. | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 6 | Reporter: | Ignacy Kasperowicz <ignacy.kasperowicz> | ||||||||||||
| Component: | anaconda | Assignee: | Anaconda Maintenance Team <anaconda-maint-list> | ||||||||||||
| Status: | CLOSED WONTFIX | QA Contact: | Release Test Team <release-test-team-automation> | ||||||||||||
| Severity: | low | Docs Contact: | |||||||||||||
| Priority: | low | ||||||||||||||
| Version: | 6.0 | CC: | dan.j.williams, ed.ciechanowski, jarema.bielanski, krzysztof.wasilewski, krzysztof.wojcik, marcin.labun | ||||||||||||
| Target Milestone: | rc | ||||||||||||||
| Target Release: | --- | ||||||||||||||
| Hardware: | x86_64 | ||||||||||||||
| OS: | Linux | ||||||||||||||
| Whiteboard: | |||||||||||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||||||||||
| Doc Text: | Story Points: | --- | |||||||||||||
| Clone Of: | Environment: | ||||||||||||||
| Last Closed: | 2010-07-22 17:57:08 UTC | Type: | --- | ||||||||||||
| Regression: | --- | Mount Type: | --- | ||||||||||||
| Documentation: | --- | CRM: | |||||||||||||
| Verified Versions: | Category: | --- | |||||||||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||||||||
| Embargoed: | |||||||||||||||
| Attachments: |
|
||||||||||||||
Created attachment 432868 [details]
Anaconda log
Created attachment 432869 [details]
Dmesg output
Created attachment 432870 [details]
Anaconda log
Created attachment 432871 [details]
Syslog
This issue has been proposed when we are only considering blocker issues in the current Red Hat Enterprise Linux release. It has been denied for the current Red Hat Enterprise Linux release. ** If you would still like this issue considered for the current release, ask your support representative to file as a blocker on your behalf. Otherwise ask that it be considered for the next Red Hat Enterprise Linux release. ** Hi, when I created IMSM (Intel) RAID 1, the installer found BIOS RAID 0MB size, ddf1_raidddf. Can you point me more closely to the place in anaconda code that that is doing RAID discovery. thanks, Marcin anaconda does not activates degraded array, therefore in case when some array member is missing the array is not activated, and the system is not installed. In the this particular reproduction method, it could happen that /dev/sda was part of array, and kickstart instructed anaconda to ignore the RAID member disk (*ignoredisk --drives=sda*). In general, all disks must be present in system, and it is suggested that IMSM metadata is the only one on the disks that are part of RAID volume on which the system is going to be installed. Lowering priority of the bug. In general, even if we are going to support installing on existing raid devices that are degraded, it's going to be up to anaconda to force their degraded assembly. As such, I'm changing the component to anaconda on this bug as I don't think mdadm has anything it needs to do. anaconda has never supported installation to degraded arrays. It's not something we intend to support. (In reply to comment #10) > anaconda has never supported installation to degraded arrays. It's not > something we intend to support. I think that the information that user receives in the situation of not activating/using degraded array is not clear; maybe anaconda could display more specific information like: list of disk that are taken into account when trying activate and the information that RAID device is not going to be activated because it is in degraded or failed state. Is it possible to display the name of BIOS raid: like DDF or IMSM, etc. Current information: "Disk sdb contains BIOS RAID metadata, but is not part of any recognized BIOS RAID sets. Ignoring disk sdb." thanks, Marcin Labun Not reproducible on RHEL6.0 Snapshot 10 x86_64. |
Created attachment 432867 [details] Storage log Description of problem: - Anaconda show error: "Disk sdb contains BIOS RAID metadata, but is not part of any recognized BIOS RAID sets. Ignoring disk sdb." - IMSM RAID volume does not start (only container is assembled but not RAID volume). - mdmon is not running (see syslog with crash info) How reproducible: Sporadically. Steps to Reproduce: 1. Clear all metadata on drives using command: mdadm --zero-superblock /dev/sd[a-f] 2. Create RAID1 in Intel OROM on 2 HDDs 3. Start kickstart installation using config file like: install nfs --server=172.28.58.30 --dir=/pxe/auto-installation key --skip lang en_US.UTF-8 keyboard us rootpw donotchange firewall --enable --ssh authconfig --enableshadow --enablemd5 selinux --enforcing timezone --utc America/New_York zerombr no ignoredisk --drives=sda bootloader --location=boot --driveorder=md127 --append="pci=nommconf rhgb quiet" clearpart --initlabel part pv.039936 --size=1 --grow --ondisk=md127 volgroup lvmvolgroup0 --pesize=32768 pv.039936 logvol / --fstype ext3 --size=37274 --name=lvm_root --vgname=lvmvolgroup0 logvol swap --fstype swap --size=2662 --name=lvm_swap --vgname=lvmvolgroup0 part /boot --fstype ext2 --size=2662 --ondisk=md127 reboot %packages @core @base emacs %post #!/bin/sh echo 1279539657 > /tmp/os_stamp 4. Installation passes and OS is bootable 5. Delete RAID in OROM and create it again (the same volume as before) 6. Start installation again with the same config file and installation process will be interrupted. In shell" # cat /proc/mdstat Personalities : [raid0] [raid1] [raid6] [raid5] [raid4] [raid10] [linear] md0 : inactive sdb[0](S) 2257 blocks super external:imsm unused devices: <none> Actual results: RAID volume is not assembled, mdmon crashes Expected results: RAID volume should be assembled properly during installation Additional info: SW: RHEL 6.0 Snapshot 7 x64 HW: DQ35JOE - ICH9 chipset Workaround: Manually delete metadata from drives using mdadm --zero-superblock command.