intermittent dns issues during container build
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
StarlingX |
Fix Released
|
Medium
|
Scott Little |
Bug Description
Brief Description
-----------------
Intermittent failures are seen when building docker images on cengn.
In the most recent case, the following images all appear to have failed on DNS lookup.
rvmc
stx-fm-subagent
stx-libvirt
stx-nova-api-proxy
stx-oidc-client
There is no clear logs on the host indicating a general networking outage. I've added a cron job to test the network and log the result.
Is it possible the default name server 8.8.8.8 (google) is having a temporary outage,
and is at fault?
Suggest adding a backup name server within the docker images.
Severity
--------
Major
Steps to Reproduce
------------------
$MY_REPO/
Expected Behavior
------------------
images are built
Actual Behavior
----------------
sometimes images fail to build
Reproducibility
---------------
intermittent, typically late night.
System Configuration
-------
N/A
Branch/Pull Time/Commit
-------
Dec 15 2020
Last Pass
---------
Dec 14 2020
Timestamp/Logs
--------------
...
fatal: unable to access 'https:/
package github.
The command '/bin/sh -c go get -d -v ./...' returned a non-zero code: 1
Command (docker) failed,
...
Could not retrieve mirrorlist http://
14: curl#6 - "Could not resolve host: mirrorlist.
The command '/bin/sh -c yum -y update' returned a non-zero code: 1
...
Test Activity
-------------
Build
Workaround
----------
Try again at a later time
tags: | added: stx.build |
stx.5.0 / medium - Results in intermittent build failues, so would be good to fix