Upgrading Fedora Atomic Host 20180515.1 to 20180722.0 failed #1471
Comments
shannara
changed the title
|
So, it looks like "Configure read-only root support" comes from Can you also paste |
|
/proc/cmdline /etc/sysconfig/readonly-root |
|
Hmm. First, we should nuke That said, that looks like the default config, and AFAICS from reading I think this is a symptom of something else; can you paste your |
|
Hi @cgwalters, This is my /etc/fstab Indicate /home and / but effective mountpoint as respectivelly /sysroot and /var/home |
|
Hmm, I can't reproduce this. I tried upgrading a cloud VM from that same version to the latest with |
|
@jlebon, thanks to try reproduce them. I use this ISO because present on my download directory, and no think to download latest. |
|
For context, from IRC, @shannara also tried the latest ISO but got issues booting: Though I just tried the latest ISO here as well in a UEFI VM with encrypted VG and a |
|
Assuming the latest ISO picked up the updated kernel, I'm guessing this is a kernel regression; maybe something going wrong with the driver for whatever disks are in use. You can test this by using |
|
Hi, Sorry to late response, off office yesterday :(, but good news, i run
:/ |
|
It sounds like there's a problem with |
|
This is status of |
cgwalters
changed the title
|
FWIW https://src.fedoraproject.org/rpms/initscripts/c/7276c52c2cf057f5a82fd73498b1c7137a615e54?branch=master split out the |
|
I hit an issue when trying out atomic workstation (silverblue). This was on a macbook pro (2015) so bare metal EFI and no dual boot. I installed from ISO (just happened to be the beta ISO I should reall try this with the officially released ISO but i'm tethered to my hotspot right now as local internet ISP is having an outage. journal.txt |
|
also, no |
|
That's quite strange...the service exits successfully according to the audit logs. But for some reason systemd (probably?) is trying to execute it again. I wonder if there's a core race condition here somewhere in systemd where ostree-remount is going in and changing the mount status, systemd detects that and decides that it hasn't completed |
|
Are you able to log on the serial console when you reach Alternatively/additionally, before running upgrade (run |
|
Humm...wait I have an idea, today our fstab generator does: But we don't have |
|
Testing out that theory now 🔬 |
|
waiting on colin to come back with the results of his test.. after that I'll try to answer @jlebon's questions |
This is standard practice for units like this; e.g. it's what `systemd-remount-fs.service` does. I think it may be part of or the whole cause for coreos/rpm-ostree#1471 I haven't reproduced the problem exactly but it seems to me that if the unit starts and is GC'd, then when systemd goes to execute a later unit it might end up restarting it. A noticeable side effect of this is that `systemctl status ostree-remount` exits with code `0` as expected.
|
At least https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=750683 confirms |
|
Yeah...I wasn't able to reproduce this in some quick testing but it's clearly a race condition per the quote:
I actually find myself wondering now what the heck a use case is for a |
This is standard practice for units like this; e.g. it's what `systemd-remount-fs.service` does. I think it may be part of or the whole cause for coreos/rpm-ostree#1471 I haven't reproduced the problem exactly but it seems to me that if the unit starts and is GC'd, then when systemd goes to execute a later unit it might end up restarting it. A noticeable side effect of this is that `systemctl status ostree-remount` exits with code `0` as expected. Closes: #1697 Approved by: jlebon
unfortunately no |
|
ostreedev/ostree#1697 seems to have solved it for me: ostreedev/ostree#1697 (comment). I just wonder why we only see this sometimes? I'd expect to see this a lot more for some reason. |
I don't know offhand; do you have a lot of mount units configured? Any unusual configuration? |
just did a kickstart install with the following partition setup and then maybe the fact that the disk is encrypted ? |
|
@shannara Do you see something like this in your log? ? |
|
Hi @cgwalters, because i reinstalling with 20180625 iso, i not previous journal of used with 20180515, but i reminding that during boot screen, few lines with mount /var/home (already mount or similar). |
|
do we think ostreedev/ostree#1697 maybe fixed the problem for this issue? should we consider closing it? |
|
Maybe? I dunno. The OP's issue had a different symptom but it could have been related, it's hard to say for sure. |
|
@shannara Any luck with a more recent FAH release? |
|
I'm going to close this since I think ostreedev/ostree#1697 fixed the problem. @shannara - please re-open if you have more info to add here or if you think the problem is not fixed. |
Host system details
Steps to reproduce it
Install Fedora Atomic 28 from 20180515.1 iso.
First issue need to create /var/cache/rpm-ostree (fix with #1366)
Second issue, after run rpm-ostree upgrade, and reboot to 20180722.0, various errors display,
waiting to mount /var/home, /boot, /boot/efi and blocked to :
"Configure read-only root support" (5mn .. / no limit)
Need to hard reboot and switch to previous ostree (20180515) that work.
The text was updated successfully, but these errors were encountered: