Bug 2001058 - systemd-udev-settle.service is deprecated. Please fix multipathd.service not to pull it in.
Summary: systemd-udev-settle.service is deprecated. Please fix multipathd.service not ...
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Fedora
Classification: Fedora
Component: device-mapper-multipath
Version: 36
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
Assignee: Ben Marzinski
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
: 2005196 (view as bug list)
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2021-09-03 16:31 UTC by Chris Murphy
Modified: 2022-11-15 19:05 UTC (History)
18 users (show)

Fixed In Version: device-mapper-multipath-0.8.9-1.fc37
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2022-11-15 19:05:22 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)
journal log (254.25 KB, text/plain)
2021-09-03 16:31 UTC, Chris Murphy
no flags Details
still happens on F36 here too (353.65 KB, image/png)
2022-04-21 15:26 UTC, Geraldo Simião
no flags Details
over two minutes waiting (143.50 KB, image/png)
2022-04-21 15:30 UTC, Geraldo Simião
no flags Details


Links
System ID Private Priority Status Summary Last Updated
Github opensvc multipath-tools issues 3 0 None None None 2021-09-10 16:46:13 UTC

Internal Links: 2182173

Description Chris Murphy 2021-09-03 16:31:08 UTC
Created attachment 1820331 [details]
journal log

Created attachment 1820331 [details]
journal log

Description of problem:

systemd still considers multipath broken by pulling in udev-settle service, which is deprecated.


Version-Release number of selected component (if applicable):
device-mapper-multipath-0.8.6-5.fc35.x86_64


How reproducible:
Always


Steps to Reproduce:
1. Boot
2.
3.

Actual results:

[    3.352525] fovo.local udevadm[609]: systemd-udev-settle.service is deprecated. Please fix multipathd.service not to pull it in.


$ systemd-analyze blame
27.623s plymouth-quit-wait.service
 3.955s fwupd.service
 1.854s systemd-udev-settle.service  <----



Expected results:

multpath should not slow down boot by 1ms let alone nearly 2 seconds. Maybe we just need to pull out multipath from Workstation edition?


Additional info:

See also bug 2001057, F35 boots 3x slower than F34.

Comment 1 Chris Murphy 2021-09-03 16:34:36 UTC
See also previous bug that closed EOL: Bug 1795013. It's a long standing problem, not a regression.

Comment 2 Chris Murphy 2021-09-03 16:36:50 UTC
'dnf remove device-mapper-multipath' only removes

Removing:
 device-mapper-multipath
Removing dependent packages:
 fcoe-utils
Removing unused dependencies:
 device-mapper-multipath-libs
 libconfig
 lldpad

Comment 3 Chris Murphy 2021-09-05 07:12:46 UTC
See also https://bugzilla.redhat.com/show_bug.cgi?id=1796437

Comment 4 Ben Marzinski 2021-09-07 17:41:30 UTC
Here is the upstream github issue for this change. https://github.com/opensvc/multipath-tools/issues/3

Comment 5 Chris Murphy 2021-09-10 16:47:31 UTC
By upstream reporting, this is a difficult problem to fix and will take some time. Setting to Rawhide.

Comment 6 Zbigniew Jędrzejewski-Szmek 2021-09-20 11:09:24 UTC
*** Bug 2005196 has been marked as a duplicate of this bug. ***

Comment 7 j7126 2022-01-01 09:06:21 UTC
Looks like this is fixed upstream now.

Comment 8 Ben Cotton 2022-02-08 21:17:47 UTC
This bug appears to have been reported against 'rawhide' during the Fedora 36 development cycle.
Changing version to 36.

Comment 9 oli 2022-04-06 11:09:09 UTC
confirm, this happens on fedora 36 beta

Comment 10 Ben Marzinski 2022-04-06 16:15:06 UTC
This will be fixed in f37, it is fixed in the current rawhide builds.

Comment 11 Geraldo Simião 2022-04-21 15:26:34 UTC
Created attachment 1874144 [details]
still happens on F36 here too

Comment 12 Geraldo Simião 2022-04-21 15:30:50 UTC
Created attachment 1874145 [details]
over two minutes waiting

Operating System: Fedora Linux 36
KDE Plasma Version: 5.24.4
KDE Frameworks Version: 5.91.0
Qt Version: 5.15.3
Kernel Version: 5.17.3-302.fc36.x86_64 (64-bit)

systemd-250.3-8.fc36.src.rpm

Comment 13 Geraldo Simião 2022-04-21 15:32:41 UTC
device-mapper-multipath-libs-0.8.7-8.fc36.x86_64
device-mapper-multipath-0.8.7-8.fc36.x86_64

Comment 14 Geraldo Simião 2022-04-21 15:47:46 UTC
applied a workarround for now (sudo systemctl mask systemd-udev-settle.service) because I don't wanna wait almost 3 minutes for a otherwise 16s boot

Comment 15 camelot 2022-05-12 00:46:07 UTC
(In reply to Geraldo Simião from comment #14)
> applied a workarround for now (sudo systemctl mask
> systemd-udev-settle.service) because I don't wanna wait almost 3 minutes for
> a otherwise 16s boot

Thx for saving 150 seconds! Seems everything still works

Comment 16 camelot 2022-08-03 13:58:18 UTC
Seems this is fixed without comment here. After unmasking (and despit deprecated warning) it exited after 3 seconds:

 systemctl status systemd-udev-settle.service
● systemd-udev-settle.service - Wait for udev To Complete Device Initialization
     Loaded: loaded (/usr/lib/systemd/system/systemd-udev-settle.service; static)
     Active: active (exited) since Wed 2022-08-03 15:46:05 CEST; 9min ago
       Docs: man:systemd-udev-settle.service(8)
    Process: 859 ExecStart=udevadm settle (code=exited, status=0/SUCCESS)
   Main PID: 859 (code=exited, status=0/SUCCESS)
        CPU: 248ms

Aug 03 15:46:03 fedora udevadm[859]: systemd-udev-settle.service is deprecated. Please fix multipathd.service not to pull it in.
Aug 03 15:46:02 fedora systemd[1]: Starting systemd-udev-settle.service - Wait for udev To Complete Device Initialization...
Aug 03 15:46:05 fedora systemd[1]: Finished systemd-udev-settle.service - Wait for udev To Complete Device Initialization.

Comment 17 Ben Marzinski 2022-08-03 14:36:26 UTC
Have you tried this in Rawhide? multipathd.service not longer pulls in systemd-udev-settle.service there. This will be fixed in f37.

Comment 18 camelot 2022-08-04 15:25:19 UTC
No, Fedora 36 aarch64 in raspberry. The same system which had the problem 3 months ago

Comment 19 Ben Marzinski 2022-11-15 19:05:22 UTC
Fixed in F37


Note You need to log in before you can comment on or make changes to this bug.