Conversation
|
Ahh doh! I did see |
| help # This message | ||
| - use '?' (question mark) as argument to get info rather than set. e.g. "max-load=?" will just print out current max-load. | ||
| `) | ||
| - use '?' (question mark) as argument to get info rather than set. e.g. "max-load=?" will just print out current max-load.\n\n`) |
There was a problem hiding this comment.
I first just dropped the newline, but then I realized it would change the output here. So I went the Fprintf() route AND had to escape the 70% r... part of the string
Yes. On Travis they take more than 1 hour to run. The tests (logic + data) are public, just running them on an internal box. |
This bothers me. This would be the third time those sepcific tests turn flaky. I fixed them twice in the past. |
|
I've pushed this branch to our git repo, which will kick the |
Hrm bummer. I was pretty sure it's not the upgrade itself that is causing the intermittent failures. I might poke at them to see if I can see what's happening on my side. Do you remember any PRs that you created as part of your fixes? |
|
@zmoazeni the repo has been updated to Go v1.14.7. Is this still required and could you update the PR if it is? 🙇 |
|
@timvaillancourt I'm going to close it unmerged. I don't have the cycles to put into it these days. |
@zmoazeni I told you so |

Related issue: #727
Description
This PR upgrades gh-ost to v1.12.4
Some notes
I ran
./script/cibuild-gh-ost-replica-testsand ran into some flakey failures in the v5.5.x MySQL line. Primarily arounddatetime-1970andmixed-charset. I'm not seeing those tests in TravisCI. Do you run those on a private box?I extended the script to allow us to pass a filter for the specific test and they intermittently seemed to be fine/fail individually (more often than not, they passed). So that has me scratching my head a little.
Let me know if you're in favor of that diff, and I'll issue a separate PR for it