Bug 1072315 - boot process hangs with Start Jobs; shutdown process hangs with Stop Jobs...
Summary: boot process hangs with Start Jobs; shutdown process hangs with Stop Jobs...
Keywords:
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: Fedora
Classification: Fedora
Component: lvm2
Version: 20
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ---
Assignee: LVM and device-mapper development team
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2014-03-04 11:18 UTC by morgan read
Modified: 2015-03-13 22:07 UTC (History)
20 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2015-03-13 22:07:00 UTC
Type: Bug
Embargoed:


Attachments (Terms of Use)
Script to gather lvm and systemd info (1.59 KB, text/plain)
2014-06-18 07:54 UTC, Peter Rajnoha
no flags Details

Description morgan read 2014-03-04 11:18:43 UTC
Description of problem:
Boot process hangs in loop with-
(1 of 3) A start job is running for dev_mapper_vg01_mm\xzdlv01_usr.local.device
(2 of 3) ditto ...05_opt.device
(3 of 3) ditto ...03_home.device
(1 of 3) ...

The only way to get out of that  is ctl-alt-del, then

Shutdown process hangs in loop with-
(1 of 9) A stop job...
...
(4 of 9) ... for LVMPV1aivCz-qiNs-ikLy-vA1q-ToYM-1Ycp-Oa81Db on /dev/dm-1
...

The only way out of that is the power switch...

The necessity of using the power switch to shutdown the system seems to indicate to me a serious problem - so I've set severity to high.

Version-Release number of selected component (if applicable):
systemd-208-14.fc20.x86_64

How reproducible:
Apparently random, but reoccurring since upgrade to f20 from f18

Steps to Reproduce:
1. upgrade to f20 from f18
2. boot f20
3.

Actual results:
Hung system with only way to un-hang system being the power switch

Expected results:
System that doesn't hang at boot or shutdown

Additional info:
Some way of escaping/ cancelling from start jobs and stop jobs is urgently needed

Comment 1 Peter Rajnoha 2014-06-18 07:54:48 UTC
Created attachment 909873 [details]
Script to gather lvm and systemd info

Could you please use attached script to gather the info - it's calling lvmdump, lsblk and systemctl/journalctl to gather info we can use for further debugging. The best would be if you could run this the next boot after the one that failed.
Then attach the archive file created by the script here in this report. Thanks.

Comment 2 Marian Csontos 2014-06-18 16:50:54 UTC
Does adding ' rd.lvm.lv=vg01/lv01_usr' (or whatever the VG and LV names are) to boot options help?

Comment 3 morgan read 2014-07-05 18:55:57 UTC
(In reply to Peter Rajnoha from comment #1)
> Created attachment 909873 [details]
> Script to gather lvm and systemd info
> 
> Could you please use attached script to gather the info - it's calling
> lvmdump, lsblk and systemctl/journalctl to gather info we can use for
> further debugging. The best would be if you could run this the next boot
> after the one that failed.
> Then attach the archive file created by the script here in this report.
> Thanks.

Just for clarity - you want that run manually under gdm once the system has booted and I've logged into user gnome account - or kicked off somehow during the boot process?

Comment 4 morgan read 2014-07-05 19:42:35 UTC
(In reply to Marian Csontos from comment #2)
> Does adding ' rd.lvm.lv=vg01/lv01_usr' (or whatever the VG and LV names are)
> to boot options help?

So that would be in /etc/default/grub here:

GRUB_CMDLINE_LINUX="rd.lvm.lv=vg01_mm/lv04_swap rd.md=0 rd.dm=0 rd.lvm.lv=vg01_mm/lv02_root rd.luks.uuid=luks-dfa915d4-6656-45f0-b771-595f2456fc39 $([ -x /usr/sbin/rhcrashkernel-param ] && /usr/sbin/rhcrashkernel-param || :) vconsole.keymap=us rd.luks.uuid=luks-81b063f2-7753-4006-b0af-30ae9d472ccd rhgb quiet"

?

Just looking at that, already there are:
rd.lvm.lv=vg01_mm/lv02_root
rd.lvm.lv=vg01_mm/lv04_swap

So, including additionally:
rd.lvm.lv=vg01_mm/lv01_usr.local
rd.lvm.lv=vg01_mm/lv03_home
rd.lvm.lv=vg01_mm/lv05_opt

Might help?

I'll try Peter's suggestions first, otherwise I might end up chassing my tail.  This is pretty intermitant.  So, it might be a while before I get another 'hit'.

Comment 5 Marian Csontos 2014-07-07 12:50:13 UTC
(In reply to morgan read from comment #4)
> (In reply to Marian Csontos from comment #2)
> > Does adding ' rd.lvm.lv=vg01/lv01_usr' (or whatever the VG and LV names are)
> > to boot options help?
> 
> So that would be in /etc/default/grub here:
> 
> GRUB_CMDLINE_LINUX="rd.lvm.lv=vg01_mm/lv04_swap rd.md=0 rd.dm=0
> rd.lvm.lv=vg01_mm/lv02_root
> rd.luks.uuid=luks-dfa915d4-6656-45f0-b771-595f2456fc39 $([ -x
> /usr/sbin/rhcrashkernel-param ] && /usr/sbin/rhcrashkernel-param || :)
> vconsole.keymap=us rd.luks.uuid=luks-81b063f2-7753-4006-b0af-30ae9d472ccd
> rhgb quiet"
> 
> ?
> 
> Just looking at that, already there are:
> rd.lvm.lv=vg01_mm/lv02_root
> rd.lvm.lv=vg01_mm/lv04_swap
> 
> So, including additionally:
> rd.lvm.lv=vg01_mm/lv01_usr.local
> rd.lvm.lv=vg01_mm/lv03_home
> rd.lvm.lv=vg01_mm/lv05_opt
> 
> Might help?

Yes, it might. Now I see you are using LUKS. What exactly is encrypted? LVs? PVs? We really need the output of scripts and commands Peter and Bryn asked for.

(In reply to morgan read from comment #3)
> (In reply to Peter Rajnoha from comment #1)
> > Created attachment 909873 [details]
> > Script to gather lvm and systemd info
> > 
> > Could you please use attached script to gather the info - it's calling
> > lvmdump, lsblk and systemctl/journalctl to gather info we can use for
> > further debugging. The best would be if you could run this the next boot
> > after the one that failed.
> > Then attach the archive file created by the script here in this report.
> > Thanks.
> 
> Just for clarity - you want that run manually under gdm once the system has
> booted and I've logged into user gnome account - or kicked off somehow
> during the boot process?

The next boot after the failed one. But running them NOW would help in meantime seeing what the layout is and may be reproduce the problem.

Comment 6 Peter Rajnoha 2014-07-07 15:19:27 UTC
(In reply to Marian Csontos from comment #5)
> > Just for clarity - you want that run manually under gdm once the system has
> > booted and I've logged into user gnome account - or kicked off somehow
> > during the boot process?
> 
> The next boot after the failed one. But running them NOW would help in
> meantime seeing what the layout is and may be reproduce the problem.

Yes, the boot after the failed one - the script I attached grabs the logs from previous (failed) boot as well as current one. We can compare it then.

Comment 7 morgan read 2015-03-13 22:07:00 UTC
This hasn't troubled me in the last 10 months, so I think it can be safely closed unless someone thinks otherwise.  Start and stop jobs always seem to resolve themselves. Thanks for listening.


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