View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0016885||CentOS-7||anaconda||public||2020-01-02 22:53||2020-01-02 22:53|
|Target Version||Fixed in Version|
|Summary||0016885: installer unable to read kickstart file from software raid partition|
|Description||I have always done major version upgrades by putting 'ks.cfg' kickstart configuration file for the new version on existing file system. With CentOS 7.7.1908 this does not seem to work if the original file systems are on software raid (md raid1) partitions. I add "inst.ks=hd:LABEL=data:/7/ks.cfg" to the kernel options line and it ends in showing dozens of some "timeout" lines and drops to emergency shell. I tried adding rd.md.uuid=xxxx option with xxxx being the UUID of the relevant md partition, to no success.|
Interestingly, when booted to normal (non-kickstart) installation, it shows all software raid partitions properly. But apparently fails to recognize them at early boot stage.
The only way I found to overcome this problem was to format one of the swap partitions (on "raw" sdaN, sdbN devices) with ext3 and to put ks.cfg there.
|Steps To Reproduce||1. CentOS 6.x system configured (/ and /export/data <labelled "data:> directories) on software raid level=1 /dev/md0 /dev/md1 partitions.|
2. Put the ks.cfg kickstart conf file for CentOS 7 on /export/data
3. Boot the 7.7.1908 installer, add "inst.ks=hd:LABEL=data:/ks.cfg" to the kernel options line
4. Installer will not find the ks.cfg, eventually dropping to the emergency shell
|Tags||installation, kickstart, software raid|