QA (public) » openQA Project (public) » openQA Infrastructure (public)

Motivation¶

See #162719 and #162725. It seems what was overlooked as part of #162719 is that w40 probably temporarily lost the connection to one of it's NVMes. Right now lsblk shows

NAME        MAJ:MIN RM   SIZE RO TYPE  MOUNTPOINTS
nvme1n1     259:0    0 476.9G  0 disk  
└─md127       9:127  0   6.3T  0 raid0 /var/lib/openqa
nvme2n1     259:1    0 476.9G  0 disk  
├─nvme2n1p1 259:2    0   512M  0 part  /boot/efi
├─nvme2n1p2 259:3    0   293G  0 part  /
…
nvme0n1     259:6    0   5.8T  0 disk  
└─md127       9:127  0   6.3T  0 raid0 /var/lib/openqa

so a RAID0 is constructed between a 500GiB and a 6TiB device which does not make much sense to me. So after all I think the approach from dheidler in #162719 was not enough as it was not addressing the real problem.

Acceptance criteria¶

• AC1: All PRG2 x86_64 OSD workers use their existing local storage more efficiently

Acceptance tests¶

• AT1-1: ssh osd sudo salt --no-color -C 'G@roles:worker and G@osarch:x86_64' cmd.run 'df -h /var/lib/openqa' shows > 1TB for all PRG2 workers

Suggestions¶

• Review what was done in #162719-15 manually, consider to change mount points using the salt states in https://gitlab.suse.de/openqa/salt-states-openqa/-/tree/master/openqa/nvme_store?ref_type=heads that prepare devices accordingly. If not feasible then apply the same either manually to all machines

Out of scope¶

Ordering any new physical storage devices