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

Issue 651126 link

Starred by 1 user

Issue metadata

Status: Archived
Owner: ----
Closed: Oct 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Bug



Sign in to add a comment

deleting and adding host to AFE at inopportune time can leave host in Running state

Project Member Reported by kevcheng@chromium.org, Sep 28 2016

Issue description

host in question chromeos6-row1-rack2-host5

mysql> SELECT * FROM afe_hosts WHERE hostname = 'chromeos6-row1-rack2-host5'\G;                         
*************************** 1. row ***************************
          id: 4903
    hostname: chromeos6-row1-rack2-host5
      locked: 0
    synch_id: NULL
      status: Running
     invalid: 1
  protection: 0
locked_by_id: NULL
   lock_time: NULL
       dirty: 1
      leased: 1
    shard_id: NULL
 lock_reason: 
1 row in set (0.00 sec)


When I readd (./atest host create ...) it, it looks like this:
mysql> SELECT * FROM afe_hosts WHERE hostname = 'chromeos6-row1-rack2-host5'\G;
*************************** 1. row ***************************
          id: 4903
    hostname: chromeos6-row1-rack2-host5
      locked: 0
    synch_id: NULL
      status: Running
     invalid: 0
  protection: 0
locked_by_id: NULL
   lock_time: NULL
       dirty: 1
      leased: 1
    shard_id: NULL
 lock_reason: 
1 row in set (0.01 sec)


And I can't do anything with it (verify, repair, job).  Plus the shard_id should be 116 (chromeos-server70 for celes).

It seems like when a host is deleted, seems like the status field should be reset to 'Repair failed' or 'Ready'.

 
Cc: xixuan@chromium.org
Xixuan resolved the issue by:

I remove the DUT from the shard again, change the leased=0 on cautotest database, then wait for some minutes, and cautotest assigns that DUT to shard 116, and shard database update itself as shard 116. Then I kick off a verify to see whether it works. 

Maybe cautotest is overload or sth happens to it. It's slow. so that the status updated on cautotest is a little bit late. After the verify complete for a while, the host's status on cautotest is just changed to 'verifying', and then after seconds, it's changed to 'ready'.
Is there more work needed on this? 

Comment 3 by autumn@chromium.org, Oct 11 2016

Status: Unconfirmed (was: Untriaged)
Project Member

Comment 4 by sheriffbot@chromium.org, Oct 12 2017

Status: Archived (was: Unconfirmed)
Issue has not been modified or commented on in the last 365 days, please re-open or file a new bug if this is still an issue.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Sign in to add a comment