rd.auto without rd.luks.[partuuid|serial|uuid] or rd.md.uuid intermittently failing (test 12)

[aafeijoo-suse]

Discovered thanks to TEST-12-RAID-DEG intermittently failing after merging #2270. The failing test is:

dracut/test/TEST-12-RAID-DEG/test.sh

Line 44 in 1843c16

client_run rd.auto || return 1

Log with rd.debug: https://github.com/dracutdevs/dracut/actions/runs/4943806249/jobs/9007358463

[aafeijoo-suse]

md0 initialized after 1mdadm -R /dev/md0 is executed? (mdraid_start.sh)

how could #2270 affect this?

ok:

[   11.096149] dracut-initqueue[742]: + _md_start /dev/md0
[   11.096489] dracut-initqueue[742]: + local _udevinfo
[   11.096808] dracut-initqueue[742]: + local _path_s
[   11.097443] dracut-initqueue[742]: + local _path_d
[   11.097785] dracut-initqueue[742]: + local _md=/dev/md0
[   11.098263] dracut-initqueue[752]: ++ udevadm info --query=property --name=/dev/md0
[   11.098977] dracut-initqueue[742]: + _udevinfo='DEVPATH=/devices/virtual/block/md0
[   11.099737] dracut-initqueue[742]: DEVNAME=/dev/md0
[   11.160865][  T755] md/raid:md0: device dm-2 operational as raid disk 1
[   11.161267][  T755] md/raid:md0: device dm-0 operational as raid disk 2
[   11.161399][  T755] md/raid:md0: raid level 5 active with 2 out of 3 devices, algorithm 2
[   11.162217][  T755] md0: detected capacity change from 0 to 155648
...
[   11.127912] dracut-initqueue[753]: ++ udevadm info -q path -n /dev/md0
[   11.128334] dracut-initqueue[742]: + _path_s=/sys//devices/virtual/block/md0/md/array_state
[   11.128824] dracut-initqueue[742]: + '[' '!' -r /sys//devices/virtual/block/md0/md/array_state ']'
[   11.190777][    T1] systemd[1]: Found device /dev/disk/by-uuid/559bb5c1-ebe5-4a97-b151-bdce99440c97.
[�[0;32m  OK  �[0m] Found device �[0;1;39m/dev/disk/by-…1-ebe5-4a97-b151-bdce99440c97�[0m.
[   11.192148][    T1] systemd[1]: Starting Cryptography Setup for luks-559bb5c1-ebe5-4a97-b151-bdce99440c97...
         Starting �[0;1;39mCryptography Setu…ebe5-4a97-b151-bdce99440c97�[0m...
[   11.132821] dracut-initqueue[754]: ++ cat /sys//devices/virtual/block/md0/md/array_state
[   11.133312] dracut-initqueue[742]: + '[' inactive '!=' inactive ']'
[   11.133813] d[   11.194612][  T785] systemd-cryptsetup[785]: Set cipher aes, mode xts-plain64, key size 512 bits for device /dev/disk/by-uuid/559bb5c1-ebe5-4a97-b151-bdce99440c97.
racut-initqueue[756]: + vinfo
[   11.134990] dracut-initqueue[756]: + read -r line
[   11.135320] dracut-initqueue[755]: + mdadm -R /dev/md0
[   11.135646] dracut-initqueue[756]: + info 'mdadm: started array /dev/md0'
[   11.136056] dracut-initqueue[756]: + echo 'mdadm: started array /dev/md0'
[   11.136473] dracut-initqueue[756]: + read -r line
[   11.136773] dracut-initqueue[756]: + '[' -n '' ']'
[   11.137224] dracut-initqueue[766]: ++ cat /sys//devices/virtual/block/md0/md/array_state
[   11.137754] dracut-initqueue[742]: + '[' read-auto = inactive ']'
[   11.138187] dracut-initqueue[742]: + _path_d=/sys//devices/virtual/block/md0/md/degraded
[   11.138757] dracut-initqueue[742]: + '[' '!' -r /sys//devices/virtual/block/md0/md/degraded ']'
[   11.139404] dracut-initqueue[742]: + :

error:

[ 59.658624] dracut-initqueue[711]: + _md_start /dev/md0
[ 59.660623] dracut-initqueue[711]: + local _udevinfo
[ 59.662692] dracut-initqueue[711]: + local _path_s
[ 59.664495] dracut-initqueue[711]: + local _path_d
[ 59.666685] dracut-initqueue[711]: + local _md=/dev/md0
[ 59.677800] dracut-initqueue[719]: ++ udevadm info --query=property --name=/dev/md0
[ 59.771436] dracut-initqueue[711]: + _udevinfo='DEVPATH=/devices/virtual/block/md0
...
[ 59.940964] dracut-initqueue[720]: ++ udevadm info -q path -n /dev/md0
[ 59.953697] dracut-initqueue[711]: + _path_s=/sys//devices/virtual/block/md0/md/array_state
[ 59.956397] dracut-initqueue[711]: + '[' '!' -r /sys//devices/virtual/block/md0/md/array_state ']'
[ 59.974645] dracut-initqueue[721]: ++ cat /sys//devices/virtual/block/md0/md/array_state
[ 59.987776] dracut-initqueue[711]: + '[' inactive '!=' inactive ']'
[ 60.003356] dracut-initqueue[722]: + mdadm -R /dev/md0
[ 60.011752] dracut-initqueue[723]: + vinfo
[ 60.013701] dracut-initqueue[723]: + read -r line
[ 60.135446] async_tx: api initialized (async)
[ 60.222103] systemd-journald[237]: Data hash table of /run/log/journal/3fb014cc9a724dff901f1ed09c698003/system.journal has a fill level at 75.1 (1537 of 2047 items, 1142784 file size, 743 bytes per hash table item), suggesting rotation.
[ 60.223719] systemd-journald[237]: /run/log/journal/3fb014cc9a724dff901f1ed09c698003/system.journal: Journal header limits reached or header out-of-date, rotating.
[ 60.383467] md/raid:md0: device sdc operational as raid disk 2
[ 60.391945] md/raid:md0: device sdb operational as raid disk 1
[ 60.405735] md/raid:md0: raid level 5 active with 2 out of 3 devices, algorithm 2
[ 60.417684] md0: detected capacity change from 0 to 155648
[ 60.384872] dracut-initqueue[723]: + info 'mdadm: started array /dev/md0'
[ 60.393716] dracut-initqueue[723]: + echo 'mdadm: started array /dev/md0'
[ 60.399576] dracut-initqueue[723]: + read -r line
[ 60.402884] dracut-initqueue[723]: + '[' -n '' ']'
[ 60.424750] dracut-initqueue[732]: ++ cat /sys//devices/virtual/block/md0/md/array_state
[ 60.443409] dracut-initqueue[711]: + '[' read-auto = inactive ']'
[ 60.445846] dracut-initqueue[711]: + _path_d=/sys//devices/virtual/block/md0/md/degraded
[ 60.447927] dracut-initqueue[711]: + '[' '!' -r /sys//devices/virtual/block/md0/md/degraded ']'
[ 60.450509] dracut-initqueue[711]: + :

[LaszloGombos]

Somewhat surprisingly the error does not seem to happen on the arch and debian or openSUSE test containers, so I am wondering if this is actually a Fedora specific issue.

[LaszloGombos]

Possibly resolved by #2454 . A possible explanation could be that right around that time Fedora transitioned 37->38 and v38 needs more RAM.

[aafeijoo-suse]

Possibly resolved by #2454 . A possible explanation could be that right around that time Fedora transitioned 37->38 and v38 needs more RAM.

Reading the logs of the failing tests, I don't think that this solves the issue.

[LaszloGombos]

Reading the logs of the failing tests, I don't think that this solves the issue.

(Sadly) I have to agree. This issue is still open for Fedora.