Slack digest for #dev - 2019-09-28

[Apache Pulsar Slack <ap...@gmail.com>]

2019-09-27 18:50:23 UTC - Addison Higham: anyone else been getting a fair number of build failures due to jenkins failing to clone? <https://builds.apache.org/job/pulsar_precommit_java8/11876/console> is an example. Probably been the 4th or 5th time I have had that happen
----
2019-09-27 18:50:49 UTC - Matteo Merli: yes, there is a combination of factors there…
----
2019-09-27 18:52:06 UTC - Matteo Merli: in some cases, there are some tests which are genuinely flaky (especially in a slower env like the ASF Jenkins). In other there are more failures event in fetching the code from git..
----
2019-09-27 18:54:16 UTC - Addison Higham: have you guys at all toyed with turning on basic retries with surefire? have wondered if that make result in a higher % of builds passing first time
----
2019-09-27 18:54:29 UTC - Addison Higham: but yeah, the git failures is the first time I have seen those in the last day or two
----
2019-09-27 18:54:43 UTC - Addison Higham: at least they fail quick though :stuck_out_tongue:
----
2019-09-27 18:59:30 UTC - Matteo Merli: We do have retries (for unit tests), though the difficult part is that in multiple cases, when the 1st execution fails it leaves the state as it is and the next retries are failing as well (eg: creating a namespace that already exists)
----