[FOLIO-987] Packer builds failing on folio-jenkins-de Created: 13/Dec/17  Updated: 12/Nov/18  Resolved: 10/Jan/18

Status: Closed
Project: FOLIO
Components: Continuous Integration
Affects versions: None
Fix versions: None

Type: Bug Priority: P2
Reporter: Wayne Schneider Assignee: Wayne Schneider
Resolution: Done Votes: 0
Labels: ci, sprint28, sprint29
Remaining Estimate: Not Specified
Time Spent: 1 day, 3 hours
Original estimate: Not Specified

Sprint:

 Description   

The folio-blackbox packer builds are failing regularly on folio-jenkins-de. It appears to be network related, but it is hard to diagnose. I get the same behavior running Vagrant on folio-jenkins-de; I don't see the issue locally.



 Comments   
Comment by Wayne Schneider [ 13/Dec/17 ]

Tried refreshing system software and rebooting, didn't seem to help.

Comment by Wayne Schneider [ 13/Dec/17 ]

Bumped up some timeouts in folio-ansible, which may have helped. Refactored the job in Jenkins to build boxes sequentially, instead of in parallel – which misses some of the point of Packer, but might help with network I/O and memory consumption on folio-jenkins-de

Comment by Wayne Schneider [ 14/Dec/17 ]

Reverted changes to folio-blackbox Jenkins job (they didn't help). Reverted docker daemon configuration change.

Current strategy – allow nightly builds to fail, build locally if necessary. Explore other bare metal options like packet.net.

Comment by John Malconian [ 15/Dec/17 ]

Reconfigured Jenkins job to use a new slave node hosted by Packet. Preliminary job runs look promising. Let's continue to monitor for a few days.

Comment by Wayne Schneider [ 10/Jan/18 ]

Have migrated jobs to folio-jenkins-slave01 on packet.net

Generated at Thu Feb 08 23:09:59 UTC 2024 using Jira 1001.0.0-SNAPSHOT#100246-sha1:7a5c50119eb0633d306e14180817ddef5e80c75d.