Bug 1192913 - Fail to boot with kernels newer than 3.17.8 - Timed out waiting for device
Summary: Fail to boot with kernels newer than 3.17.8 - Timed out waiting for device
Keywords:
Status: CLOSED WORKSFORME
Alias: None
Product: Fedora
Classification: Fedora
Component: systemd
Version: 21
Hardware: x86_64
OS: Linux
unspecified
urgent
Target Milestone: ---
Assignee: systemd-maint
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2015-02-16 08:16 UTC by Dietrich
Modified: 2015-04-24 12:19 UTC (History)
17 users (show)

Fixed In Version:
Clone Of:
Environment:
Last Closed: 2015-04-24 12:19:52 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
the screen that the boot is stuck with. (166.47 KB, image/jpeg)
2015-02-16 08:18 UTC, Dietrich
no flags Details
the output from lspci (5.08 KB, text/plain)
2015-02-16 08:19 UTC, Dietrich
no flags Details
the output from mount (1.47 KB, text/plain)
2015-02-16 08:20 UTC, Dietrich
no flags Details
the output from mount (2.70 KB, text/plain)
2015-02-16 08:21 UTC, Dietrich
no flags Details
the output of blkid (1.47 KB, text/plain)
2015-02-16 08:22 UTC, Dietrich
no flags Details
Second timeout image after an even longer period of waiting (2.30 MB, image/jpeg)
2015-02-16 11:51 UTC, Dietrich
no flags Details
output of journalctl -b (497.36 KB, text/plain)
2015-02-16 12:19 UTC, Dietrich
no flags Details

Description Dietrich 2015-02-16 08:16:18 UTC
User-Agent:       Mozilla/5.0 (X11; Linux x86_64; rv:35.0) Gecko/20100101 Firefox/35.0
Build Identifier: 

When booting with the most recent kernels my system won't boot! I don't get any emergency shell or any interaction possibility which is why this bug report is not as complete as I'd like it to be.

On Boot I get the usual Messages:
[OK] startet ...
[OK] ...

I didn't find anything wrong in those messages until I get 10 services that wait. The timeout is 1.30 min and after the timeout the screen is stuck at the picture that I will upload. listing the failed services.

Reproducible: Always

Steps to Reproduce:
1. Boot the system with newer kernel than 3.17.8
2. wait for timeout
3. observe screen as in screenshot
Actual Results:  
bootup is stuck at the screenshot

Expected Results:  
smoothly bootup into gnome login screen.

I tried to install the newer kernel 3.19 from fedora 22 but no improvement.

I will attach various outputs from different commands so that you might be able to get my disk layouts.

Comment 1 Dietrich 2015-02-16 08:18:02 UTC
Created attachment 992065 [details]
the screen that the boot is stuck with.

Comment 2 Dietrich 2015-02-16 08:19:09 UTC
Created attachment 992066 [details]
the output from lspci

Comment 3 Dietrich 2015-02-16 08:20:14 UTC
Created attachment 992067 [details]
the output from mount

these are the outputs on the working 3.17 kernel!

Comment 4 Dietrich 2015-02-16 08:21:25 UTC
Created attachment 992068 [details]
the output from mount

these are the outputs with the working kernel 3.17

Comment 5 Dietrich 2015-02-16 08:22:16 UTC
Created attachment 992069 [details]
the output of blkid

Comment 6 Dietrich 2015-02-16 11:51:05 UTC
Created attachment 992151 [details]
Second timeout image after an even longer period of waiting

I noticed that after waiting for some more minutes I get more messages on my screen. They are all the same... so there are some more than in that image but I only posted the first ones. I didn't find any different messages.

Comment 7 Dietrich 2015-02-16 12:14:05 UTC
Ok I got to boot my system by disabling many services I didn't need anyways and by removing the readahead file. I do not really know which of the changes worked around the problem but I guess it was disabling iscsi.

  547  sudo systemctl disable lvm2-monitor.service
  548  sudo systemctl disable lvm2-lvmetad.service
  549  sudo systemctl disable sendmail.service
  550  sudo systemctl disable sm-client.service
  551  sudo systemctl disable bluetooth.service
  552  sudo systemctl disable plymouth-read-write.service
  553  sudo systemctl disable plymouth-start.service
  554  sudo systemctl disable dmraid-activation.service
  555  sudo systemctl disable livesys-late.service
  556  sudo systemctl disable livesys.service
  557  systemctl disable systemd-readahead-collect systemd-readahead-replay
[...]
  563  sudo mv /.readahead /noreadahead
[...]
  567  systemctl disable iscsid.socket
  568  sudo systemctl disable iscsiuio.socket

I will attach a log from the successful boot for completeness sake.

Comment 8 Dietrich 2015-02-16 12:19:26 UTC
Created attachment 992157 [details]
output of journalctl -b

After deactivating several services that are not needed anyways the system boots again. The log is from the first successful boot on the new kernel. I will leave it up to someone else to judge if the issue I encountered needs more investigation or is ok to be closed sorry for the time I made you guys spent on this issue.

Comment 9 afreof 2015-04-04 08:13:35 UTC
Disabling all units related to readahead seems to solve the problem.
The issue occurred on three out of five machines. All of them have a btrfs root file system.
Readahead has been removed from systemd (release 219). Hence, the problem is already solved upstream. I would like to see a systemd patch for Fedora 21 which disables readahead as well.

Comment 10 Linus Walleij 2015-04-21 18:53:03 UTC
Are you using BTRFS for any of the default mounts? (Check /etc/fstab)

If that is the case you are seeing bug 1187106.

Comment 11 Jan Synacek 2015-04-24 06:24:37 UTC
In https://bugzilla.redhat.com/attachment.cgi?id=992068, I can see:
/dev/sdc3 on / type btrfs (rw,relatime,seclabel,ssd,space_cache)

Comment 12 Dietrich 2015-04-24 12:19:52 UTC
Thank you for the input. though I'm not entirely sure whether it is actually a duplicate of said bug 1187106 it is working for me now and seems to be only related to stale configs etc. I close the bug with "works for me" for now.


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