New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 860034 link

Starred by 3 users

Issue metadata

Status: Fixed
Owner:
Closed: Jul 16
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 2
Type: Bug



Sign in to add a comment

minijail0 tries to bind-mount /var/log/memd before memd creates it

Project Member Reported by semenzato@chromium.org, Jul 3

Issue description

It's pretty obvious from the code (platform2/src/metrics/memd/init/memd.conf) but here's a log:


2018-07-02T14:04:24.602865-07:00 WARNING memd[2306]: libminijail[2]: stat(/var/log/memd) failed: No such file or directory
2018-07-02T14:04:24.602879-07:00 WARNING memd[2306]: libminijail[2]: creating mount target '/var/empty/var/log/memd' failed
2018-07-02T14:04:24.608523-07:00 ERR memd[2306]: libminijail[2]: mount_one failed: No such file or directory
2018-07-02T14:04:24.627206-07:00 INFO crash_reporter[2307]: libminijail[2307]: mount /dev/log -> /dev/log type ''
2018-07-02T14:04:24.668937-07:00 WARNING crash_reporter[2307]: [user] Received crash notification for memd[2306] sig 6, user 0 (developer build - not testing - always dumping)
2018-07-02T14:04:24.675826-07:00 INFO crash_reporter[2307]: State of crashed process [2306]: S (sleeping)
2018-07-02T14:04:24.712276-07:00 INFO crash_reporter[2307]: Stored minidump to /var/spool/crash/memd.20180702.140424.2306.dmp
2018-07-02T14:04:24.712657-07:00 INFO crash_reporter[2307]: Leaving core file at /var/spool/crash/memd.20180702.140424.2306.core due to developer image
2018-07-02T14:04:24.713219-07:00 WARNING crash_reporter[2307]: [ARC] Received crash notification for memd[2306] sig 6, user 0 (ignoring - crash origin is not ARC)
2018-07-02T14:04:24.725435-07:00 WARNING kernel: [   18.221531] init: memd main process (2305) terminated with status 254

I'll send a CL shortly and we can discuss the solution in there.
 
Cc: puthik@chromium.org sonnyrao@chromium.org
The CL Luigi mentioned is here:
https://chromium-review.googlesource.com/c/chromiumos/platform2/+/1125267
Project Member

Comment 2 by bugdroid1@chromium.org, Jul 11

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform2/+/e6f46c840574f04f9407da8d991a1c1b8cc1476f

commit e6f46c840574f04f9407da8d991a1c1b8cc1476f
Author: Luigi Semenzato <semenzato@chromium.org>
Date: Wed Jul 11 15:40:15 2018

metrics: memd: create /var/log/memd at boot

Chicken and egg problem: memd creates /var/log/memd if it doesn't
exist, but minijail comes first.

BUG= chromium:860034 
TEST=rm -rf /var/log/memd, start memd before and after the fix

Change-Id: I10570367583554ce88db36650146273c1e672351
Reviewed-on: https://chromium-review.googlesource.com/1125267
Commit-Ready: Luigi Semenzato <semenzato@chromium.org>
Tested-by: Luigi Semenzato <semenzato@chromium.org>
Reviewed-by: Sonny Rao <sonnyrao@chromium.org>

[modify] https://crrev.com/e6f46c840574f04f9407da8d991a1c1b8cc1476f/metrics/memd/init/memd.conf

Labels: ReleaseBlock-Beta
most likely yes.  unfortunately the fix landed in 10867.0.0 but 10866.0.0 was pushed to dev.

crash/1d72048cb25f98c9
crash/5e087d63c02b581f
crash/2ab60e006dcbdba7
crash/f7981a5d73e9fb4d
crash/76e5d5fcc07006c3
crash/8dde0050de70e638
crash/fd0ff3e25f9fc902
crash/1093761a2c58e2e5
crash/215e0222b1c684f8
should we turn off memd for R69 ?  we're pretty close to the branch and clearly it hasn't seen much testing so far.  we can turn it back on once R69 branches.
Labels: M-69
Moving to RBD as this is suspected to be causing login loop crash ( issue 863261 )
Labels: -ReleaseBlock-Beta ReleaseBlock-Dev
Luigi is still OOO until monday I think, but I'd say we should disable it if you think it's causing problems
i've posted https://chromium-review.googlesource.com/1136874 to turn off memd for R69
ok thanks, yeah we can look at it again for 70
Can RBD be removed as it appears the target has changed?
I was thinking we can do that once the CL in #10 lands?
Project Member

Comment 14 by bugdroid1@chromium.org, Jul 15

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/f4efe08ab35905cde61ec77c1f47b85e7864b40d

commit f4efe08ab35905cde61ec77c1f47b85e7864b40d
Author: Mike Frysinger <vapier@chromium.org>
Date: Sun Jul 15 01:11:43 2018

memd: disable for R69

It's pretty late in the branch cycle, so pull memd out as it hasn't
had enough testing just yet.  We can re-enable for R70.

BUG= chromium:860034 ,chromium:729335
TEST=precq passes

Change-Id: Ib8f97c57689ce5d14bf24120643df419412dc7e0
Reviewed-on: https://chromium-review.googlesource.com/1136874
Commit-Ready: Mike Frysinger <vapier@chromium.org>
Tested-by: Mike Frysinger <vapier@chromium.org>
Reviewed-by: Sonny Rao <sonnyrao@chromium.org>

[modify] https://crrev.com/f4efe08ab35905cde61ec77c1f47b85e7864b40d/profiles/targets/chromeos/make.defaults

Cc: cindyb@chromium.org kbleicher@chromium.org
Can the status be transitioned from started?  Also, is this associated with crbug/863261, crbug/860741, and/or crbug/861995?  Can we tag the DUPs, dependencies, etc.?

Thanks.  Trying to close dev blockers as a priority.  Thanks
Status: Fixed (was: Started)
it's been fixed for a while
I have just seen this crash https://crash.corp.google.com/browse?q=reportid=%27123128db7fdb1992%27 on following build:

Google Chrome	69.0.3486.0 (Official Build) dev (64-bit)
Platform	10866.1.0 (Official Build) dev-channel eve


yeah, the dev channel picked up the buggy state before this CL landed, and we haven't had a new dev channel since.  10866.1.0 is known to be broken.
Issue 868759 has been merged into this issue.
I think that the crash report only indicates that the service crashed.  The root cause is almost surely different, because on this particular bug the crash happened immediately after boot, but here it happens about 2 hours into a session.  I will file a separate bug, thanks.
Created crbug/888700.  Thanks

Sign in to add a comment