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

Issue 597680 link

Starred by 2 users

Issue metadata

Status: Verified
Owner:
Closed: Jun 2016
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: 1
Type: ----

Blocking:
issue 582646



Sign in to add a comment

Devserver may fail staging when using SSH

Project Member Reported by xixuan@chromium.org, Mar 24 2016

Issue description

SSH can be aborted by devserver, which causes failure in CQ like:

Running 'ssh 172.17.40.17 'curl "http://172.17.40.17:8082/is_staged?artifacts=full_payload,stateful,autotest_packages&files=&archive_url=gs://chromeos-image-archive/x86-alex-paladin/R51-8104.0.0-rc2"''

03/23 22:52:10.740 DEBUG|        base_utils:0177| Running 'ssh 172.17.40.17 'curl "http://172.17.40.17:8082/is_staged?artifacts=full_payload,stateful,autotest_packages&files=&archive_url=gs://chromeos-image-archive/x86-alex-paladin/R51-8104.0.0-rc2"''

03/23 22:52:17.009 DEBUG|        base_utils:0177| Running 'ssh 172.17.40.17 'curl "http://172.17.40.17:8082/is_staged?artifacts=full_payload,stateful,autotest_packages&files=&archive_url=gs://chromeos-image-archive/x86-alex-paladin/R51-8104.0.0-rc2"''

03/23 22:52:17.176 DEBUG|        dev_server:0728| Error occurred with exit_code 255 when executing thessh call:

ssh_exchange_identification: Connection closed by remote host


Current mechanism don't retry ssh connection if we enter an ssh 'CmdError'. That may be an issue.
 

Comment 1 by fdeng@chromium.org, Mar 24 2016

imo we should retry the CmdError here. 

Project Member

Comment 2 by bugdroid1@chromium.org, Mar 29 2016

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/third_party/autotest/+/476408397bec660680d7d4f347ead1eb7dc21aa4

commit 476408397bec660680d7d4f347ead1eb7dc21aa4
Author: xixuan <xixuan@google.com>
Date: Thu Mar 24 17:33:16 2016

Autotest: Make Devserver retry ssh connection, and Add stats for ssh
failures.

When trying to SSH to a devserver, this SSH connection could be aborted
by the devserver itself, which causes staging failure.

This CL add 'retry' to all ssh connections to devserver, and add stats
for ssh failures: ssh_dev_server_failure, and ssh_tunnel_failure, to
estimate the stabality of ssh-based approaches.

BUG= chromium:597680 
TEST=Ran dev_server_unittest to ensure error.CmdError causes retry. Turn
on the enable_ssh switch to test that ssh to devserver & servo still works.

Change-Id: I4b9e6092d8989fb2ecb770367c5869042e8e5740
Reviewed-on: https://chromium-review.googlesource.com/334770
Reviewed-by: Xixuan Wu <xixuan@chromium.org>
Tested-by: Xixuan Wu <xixuan@chromium.org>
Commit-Queue: Xixuan Wu <xixuan@chromium.org>

[modify] https://crrev.com/476408397bec660680d7d4f347ead1eb7dc21aa4/server/hosts/rpc_server_tracker.py
[modify] https://crrev.com/476408397bec660680d7d4f347ead1eb7dc21aa4/client/common_lib/cros/dev_server.py
[modify] https://crrev.com/476408397bec660680d7d4f347ead1eb7dc21aa4/client/common_lib/cros/dev_server_unittest.py

Comment 3 by benhenry@google.com, Apr 26 2016

Components: Infra>Client>ChromeOS
Labels: -Infra-ChromeOS
Status: Fixed (was: Assigned)
Closing... please feel free to reopen if its not fixed.
Status: Verified (was: Fixed)

Sign in to add a comment