Please prioritize and assign.

Hi Juan,

Was the device was auto-launching the kiosk app? How long did you want before it turned red?

If you waited >30 minutes and did enable auto-launch please open a b/ and include:

* Device Serial
* Domain
* Date/Time this was tested and the network it was on?

Ping me with that ID when you have it so we can investigate as this is controlled on the server side.

Thanks

Hi Matt
-the device is in the lab autolaunching and currently running App Stratosmedia (accessible via ssh root@chromeos1-row1-rack5-host2) 
-device is enrolled in the croste.tv/longevity domain.
-The device has been red since I checked this morning.
from the vpd -l got these SNs:
"mlb_serial_number"="QTFMNX0849005899"
"serial_number"="CN0KNJ204864349J0551A00"

I checked the log, and the device is sending heartbeats as expected, hence this is a server issue. Assigned the bug to me.

I manually fixed this issue for this particular device. Meanwhile, filed a server bug (https://b.corp.google.com/u/0/issues/27796962) to track the issue.

Closing the client bug for now.

Hi Jin
could you please elaborate how you fixed this issue? Thanks

Sure. I modified the cache entry on the server side. Before the fix, the cache entry stores the last_device_status = ONLINE in the memcache entry, which is correct, since heartbeats are received every 2 minutes. This would make status monitor job NOT update the DDS record, because it thinks the device status didn't change (old and new are both ONLINE).

However, in DDS, the status is OFFLINE, probably because a write to DDS that updates the device status from OFFLINE to ONLINE failed. The device got stuck in this case, hence I manually modify the cache entry and changed the last_device_status = OFFLINE. After that, status monitor would update the DDS record because the device status in cache got toggled.

This is a temporary fix. I filed a server bug to track this issue and would fix it for good.

Thanks Jin for looking into it, it makes sense. This is a temporary fix though, our test is a continuous 24/7 test, how do we know when we see a red dot for health status that the same problem is repeating again, how do we diagnose it? or should we just file a bug every time that the device appears to be offline? Thanks

Hi Juan,

I'll be back to office next Tue, and will fix this issue then. From now to 3/29, you can file a bug and assigned to me when a device is believed to be online, but stays offline for more than 30 min. In the bug, please tell me the serial number and the domain name.

Tricky device shows off line status in CPanel

Deleted: log-032316-090614.tar.gz 1.9 MB

log-032316-090614.tar.gz 1.9 MB Download

Fixed per https://b.corp.google.com/u/0/issues/27796962.

bug verified. Device shows Device status in CPanel (green). Also is displayed "memory usage", "CPU utilization", "Disk Space", "App info", Screen captures and system logs also work