[
https://issues.apache.org/jira/browse/SOLR-10109?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15860300#comment-15860300
]
Mark Miller edited comment on SOLR-10109 at 2/9/17 10:11 PM:
-------------------------------------------------------------
This class is my old friend (5 years ago going by AutoCommitTest). I have
dozens of tiny commits trying to straighten out it's odd wonders by simply
pushing on timeout limits.
It's always been a fail happy test, though I think sometimes it's been better
and other times worse. It's trying too hard to test shortish timings I think.
It also looks like SOLR-10110 really made what was going on in this test even
worse.
Anyway, for the last half a day I battled it from 40-50% fails in 30 to 1 fail
in 30 to 1 fail in 60 to a couple fails in 200 until now when I've gotten out a
few 200 run beastings with no fails.
Does not mean they don't exist, but I'm praying there is some improvement.
was (Author: [email protected]):
This class is my old friend. I have dozens of tiny commits trying to straighten
out it's odd wonders by simplifying pushing on timeout limts.
It's always been a fail happy test, though I think sometimes it's been better
and other times worse. It's trying too hard to test shortish timings I think.
It also looks like SOLR-10110 really made what was going on in this test even
worse.
Anyway, for the last half a day I battled it from 40-50% fails in 30 to 1 fail
in 30 to 1 fail in 60 to a couple fails in 200 until now when I've gotten out a
few 200 run beastings with no fails.
Does not mean they don't exist, but I'm praying there is some improvement.
> SoftAutoCommitTest is too fragile.
> ----------------------------------
>
> Key: SOLR-10109
> URL: https://issues.apache.org/jira/browse/SOLR-10109
> Project: Solr
> Issue Type: Test
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Mark Miller
> Assignee: Mark Miller
>
> In my beast test runs it has failed 7% of the time, then 3% of the time, then
> when I started using hardware with 8 cores instead of 16, 40% of the time
> (unless something else also shifted). That gives it a test rating of 'screwy'.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
