Project

General

Profile

action #89551

Updated by mkittler about 3 years ago

### problem 
 * The NFS mount for `/var/lib/openqa/share` fails when booting. This has been reproduced on `openqaworker13` (see #88900), `openqaworker2` (see #89551#note-5) #88900) and possibly `openqaworker-arm-2` (see #75016). 
 * Technically all workers might be affected because the problem is quite generic: The systemd service for the NFS mount does not wait until the ethernet connection is established. 

 ### impact 
 The mounting is actually re-attempted automatically after a few minutes so it shouldn't be a big deal. However, since the systemd unit for the mount stays failed for a few minutes false alerts are triggered which should be prevented. 

 ### acceptance criteria 
 * AC1: No false alerts are triggered if the NFS mountpoint fails just for a few minutes after booting. 

 ### notes 
 * The dependency to `network-online.target` is correct. However, `network-online.target` does not really ensure the ethernet connection is up in our setup. This is pretty clear from the logs mentioned in #88900#note-5. 
 * To configure `network-online.target` correctly further configuration would be required. 
     * The official documentation only mentions how to do it for NetworkManager and systermd-networkd: https://www.freedesktop.org/wiki/Software/systemd/NetworkTarget/#cutthecraphowdoimakesurethatmyservicestartsafterthenetworkisreallyonline 
     * Likely it doesn't make much sense relying on `network-online.target` anyways and just cope with the fact that the mountpoint might fail for a few minutes.

Back