Unexpected unenrollment of May Flowers |
||||||||||||
Issue descriptionShort version: - Chromebox had a flashing white light power button - Pressing once shows the login screen on MIMO (indicating the unenrollment) - Device was manually enrolled and then functioned fine We should look at the device's old logs to look for unenrollment potential issues (disk full, etc) ---------- Based on the investigation in b/64203855 we are worried that the existence the following line may indicate an ARC++ event happening on the device 2017-08-21T18:51:34.068686+00:00 ERR kernel: [ 1.407992] EXT4-fs (dm-0): couldn't mount as ext3 due to feature ---------- 1) We need to find out if this device was supposed to get ARC++ 2) And even if it was supposed to get, we may want to think about blocking ARC++ on CFM devices until we have full blown testing completed.
,
Aug 23 2017
,
Aug 23 2017
Looking at the context of those logs, I'm not super confident that EXT4-fs line is relevant. - Most operating systems go through automatic FS type detection - And its possible that in this case ext2 was mounted by ext4 subsystem - There is no evidence that it would have caused an issue just because of the presence of this line - We need to check when the ext4 subsystem was introduced. It could predate M61. 2017-08-21T18:51:34.068686+00:00 ERR kernel: [ 1.407992] EXT4-fs (dm-0): couldn't mount as ext3 due to feature incompatibilities 2017-08-21T18:51:34.068687+00:00 INFO kernel: [ 1.408193] EXT4-fs (dm-0): mounting ext2 file system using the ext4 subsystem 2017-08-21T18:51:34.068689+00:00 INFO kernel: [ 1.409495] EXT4-fs (dm-0): mounted filesystem without journal. Opts: (null)
,
Aug 23 2017
Approving merge to M61 Chrome OS.
,
Aug 24 2017
,
Aug 24 2017
Hi Gwendal, can you check if the error message is of concern or not? We are seeing data lost causing unenrollment of CFM.
,
Aug 24 2017
#3, this is expected: dm-0 (root partition) is using ext2, but we use ext4 as the driver in the kernel. It tries ext3 first and rollback to ext2.
,
Aug 25 2017
Hi Bin, We are trying to chase down enterprise enrollment issue, can you tell us who is the best engineer to discuss this?
,
Aug 28 2017
This issue has been approved for a merge. Please merge the fix to any appropriate branches as soon as possible! If all merges have been completed, please remove any remaining Merge-Approved labels from this issue. Thanks for your time! To disable nags, add the Disable-Nags label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Aug 28 2017
,
Aug 28 2017
,
Aug 28 2017
,
Aug 28 2017
i'm removing the launch blocker - imo, we have rooms that exhibit unexpected unenrollment issues already and we have no specific data that this is a specific regression other issues about losing stateful information seems to have been cleared up in other bugs (ie changing domains so changing file names of where data is stored)
,
Oct 10 2017
,
Dec 4 2017
|
||||||||||||
►
Sign in to add a comment |
||||||||||||
Comment 1 by choonc@google.com
, Aug 23 2017