Description of problem:
multipathd logs during boot at error level:
Sep 21 15:51:51 multipathd: /etc/multipath.conf does not exist, blacklisting all devices.
Sep 21 15:51:51 multipathd: You can run "/sbin/mpathconf --enable" to create
Sep 21 15:51:51 multipathd: /etc/multipath.conf. See man mpathconf(8) for more details
Not having any multipath devices is completely mundane, the service should not report this as an error. See https://fedoraproject.org/wiki/Packaging:DefaultServices#Must_not_fail_under_normal_operating_conditions
Version-Release number of selected component (if applicable):
A single log line at info level, "/etc/multipath.conf does not exist, multipathd exiting" or something like that.
This only happens when manually running multipathd, correct? Manually starting multipathd without a configuration file is not completely mundade. It's something that should never happen.
Systemd refuses to start the service because of
Which will keep you from ever seeing this message. So, unless I'm missing something, you have to manually run multipathd to see this, which you should not be doing anyway. You should start it via systemd. I would argue that if someone thinks that they have correctly started multipathd but they haven't, that's important information. Is there some situation during boot where something is trying to start multipathd not through systemd?
I omitted an important bit of information: this is on silverblue (F29-beta5 amd64 silverblue image under qemu). This might or might not be relevant.
> This only happens when manually running multipathd, correct?
No, I'm not starting anything manually. This error message is from the logs from a default boot after installation.
The .service has ConditionPathExists=/etc/multipath.conf, but .socket does not. I guess the service is started when something connects to the socket.
Created attachment 1520673 [details]
screenshot from qemu boot
Can we please get this fixed? When booting a Workstation-live image (and xfce, and probably others), this is prominently visible during but and *very* ugly. People are putting a lot of effort into polish, and stuff like this completely defeats it.
Please see attached screenshot from
qemu-system-x86_64 -machine accel=kvm -smp 2 -m 2048 -drive if=pflash,format=raw,readonly,file=/usr/share/edk2/ovmf/OVMF_CODE.fd -cdrom Workstation-Live-x86_64-Rawhide-20190114.n.0.iso
device-mapper-multipath-0.7.7-6.gitef6d98b.fc29 has been submitted as an update to Fedora 29. https://bodhi.fedoraproject.org/updates/FEDORA-2019-e969fe2844
device-mapper-multipath-0.7.7-6.gitef6d98b.fc29 has been pushed to the Fedora 29 testing repository. If problems still persist, please make note of it in this bug report.
See https://fedoraproject.org/wiki/QA:Updates_Testing for
instructions on how to install test updates.
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2019-e969fe2844
device-mapper-multipath-0.7.7-6.gitef6d98b.fc29 has been pushed to the Fedora 29 stable repository. If problems still persist, please make note of it in this bug report.