New issue
Advanced search Search tips

Issue 891200 link

Starred by 1 user

Issue metadata

Status: Fixed
Owner:
Closed: Oct 4
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 1
Type: Bug

Blocked on:
issue 888744



Sign in to add a comment

coral-paladin HWTest failed at build_RootFileSystemSize (free bytes less than required)

Project Member Reported by yamaguchi@chromium.org, Oct 2

Issue description

coral-paladin #4345
https://cros-goldeneye.corp.google.com/chromeos/healthmonitoring/buildDetails?buildbucketId=8933811125720017568


https://luci-logdog.appspot.com/v/?s=chromeos/bb/chromeos/coral-paladin/4345/+/recipes/steps/HWTest__bvt-inline___astronaut_/0/stdout

======
  graphics_dEQP.bvt                       [ PASSED ]
  login_RetrieveActiveSessions            [ PASSED ]
  build_RootFilesystemSize                [ FAILED ]
  build_RootFilesystemSize                  FAIL: 7970816 bytes free is less than the 11534336 required.
  build_RootFilesystemSize                  retry_count: 2
======

Not sure if this is transitive issue or not.
#4344 was success.
#4343 was by a different issue (AC power is not plugged in).
 
Labels: Restrict-View-Google
I've -1'd this change in CQ tentatively.
https://chrome-internal-review.googlesource.com/c/chromeos/overlays/overlay-coral-private/+/679540
Although not 100% sure that it caused error, the change is the only one specific to coral and says "firmware uprev".
Labels: -Restrict-View-Google
Owner: dgarr...@chromium.org
Status: Assigned (was: Available)
Dan, will you take a look? My hypothesis was that the change above didn't match with the current config of the DUT, or the builder config, so removed it from CQ tentatively to see how it goes.
It seems very strange that this is a lab test, but the condition it tests for is:

        # Fail if we are running out of free space on rootfs (11 MiB or
        # 2% free space).
        required_free_space = min(11 * 1024 * 1024, used * 0.02)

So, any CL that increases rootfs usage could be to blame.
What architecture is coral?
And for the general case of this, see https://crbug.com/888744
Blockedon: 888744
I think it was just this CL [1] blowing our space. Do we have any other action, or is this just a dupe of bug 888744?

Anyway, if that *is* true, it seems anyone developing coral derivatives is in for a bad time...

[1] https://chrome-internal-review.googlesource.com/c/chromeos/overlays/overlay-coral-private/+/679540
>  build_RootFilesystemSize                  FAIL: 7970816 bytes free is less than the 11534336 required.

My estimation with https://chrome-internal-review.googlesource.com/c/chromeos/overlays/overlay-coral-private/+/679540 the size of chromeos-firmwareupdate will be increased by 4~5MB, which is larger than the delta of ~3.4MB (i.e. 11534336 - 7970816)

So the size of rootfs looks to be the issue?
Experimenting with https://chromium-review.googlesource.com/c/chromiumos/overlays/board-overlays/+/1260243 to increase rootfs size to see if that works.
Cc: akhouderchah@chromium.org benchan@chromium.org
I think benchan is already doing this for his own sake:

https://chromium-review.googlesource.com/c/chromiumos/overlays/board-overlays/+/1255822

Quick, land your coral bloat!!
Status: Fixed (was: Assigned)
Another CL https://chromium-review.googlesource.com/c/chromiumos/overlays/board-overlays/+/1255822 landed already. As a result, https://chrome-internal-review.googlesource.com/c/chromeos/overlays/overlay-coral-private/+/679540 passed the CQ and has been merged.

Mark it fixed for now. Please re-open if in need.

Thanks.
Project Member

Comment 12 by bugdroid1@chromium.org, Oct 7

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/board-overlays/+/aac1c0ce1b330cdf394acb1c8c36b5d7fd34f872

commit aac1c0ce1b330cdf394acb1c8c36b5d7fd34f872
Author: Ben Chan <benchan@chromium.org>
Date: Sun Oct 07 23:34:16 2018

baseboard-coral: increase rootfs size by 20MB

After including additional modem firmware, the space left in rootfs
becomes too small to pass the build_RootFilesystemSize test. This CL
includes the rootfs again to give us some room to breathe.

BUG=b:117099731
BUG= chromium:891200 
TEST=build_RootFilesystemSize passes with additional modem firmware included.

Change-Id: I69b66dcd5c6f876e3021c8d8ce104389eabe207f
Reviewed-on: https://chromium-review.googlesource.com/c/1266915
Tested-by: Ben Chan <benchan@chromium.org>
Reviewed-by: Ben Chan <benchan@chromium.org>

[modify] https://crrev.com/aac1c0ce1b330cdf394acb1c8c36b5d7fd34f872/baseboard-coral/scripts/disk_layout.json

Project Member

Comment 13 by bugdroid1@chromium.org, Oct 10

Labels: merge-merged-release-R70-11021.B
The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/board-overlays/+/b1565702031cc8d4b5fa2ee7973ddbf99bdd2c0c

commit b1565702031cc8d4b5fa2ee7973ddbf99bdd2c0c
Author: Ben Chan <benchan@chromium.org>
Date: Wed Oct 10 16:45:43 2018

baseboard-coral: increase rootfs size by 20MB

After including additional modem firmware, the space left in rootfs
becomes too small to pass the build_RootFilesystemSize test. This CL
includes the rootfs again to give us some room to breathe.

BUG=b:117099731
BUG= chromium:891200 
TEST=build_RootFilesystemSize passes with additional modem firmware included.

Change-Id: I69b66dcd5c6f876e3021c8d8ce104389eabe207f
Reviewed-on: https://chromium-review.googlesource.com/c/1266915
Tested-by: Ben Chan <benchan@chromium.org>
Reviewed-by: Ben Chan <benchan@chromium.org>
(cherry picked from commit aac1c0ce1b330cdf394acb1c8c36b5d7fd34f872)
Reviewed-on: https://chromium-review.googlesource.com/c/1273619
Commit-Queue: Ben Chan <benchan@chromium.org>

[modify] https://crrev.com/b1565702031cc8d4b5fa2ee7973ddbf99bdd2c0c/baseboard-coral/scripts/disk_layout.json

Sign in to add a comment