Re: [v15 beta] pg_upgrade failed if earlier executed with -c switch
Michael Paquier <michael@paquier.xyz>
From: Michael Paquier <michael@paquier.xyz>
To: Justin Pryzby <pryzby@telsasoft.com>
Cc: Daniel Gustafsson <daniel@yesql.se>, Tom Lane <tgl@sss.pgh.pa.us>, tushar <tushar.ahuja@enterprisedb.com>, pgsql-hackers@postgresql.org, Andrew Dunstan <andrew@dunslane.net>
Date: 2022-06-09T00:53:36Z
Lists: pgsql-hackers
Attachments
- 0001-Update-TestUpgrade.pm-to-grab-for-any-existing-log-f.patch (text/x-diff) patch 0001
On Wed, Jun 08, 2022 at 04:13:37PM -0500, Justin Pryzby wrote: > On Wed, Jun 08, 2022 at 10:55:29AM +0900, Michael Paquier wrote: >> And applied, to take care of this open item. > > Shouldn't this wait for the buildfarm to be updated again ? The TAP logic is able to find any logs by itself on failure, so what would be impacted is the case of the tests running pg_upgrade via the past route in TestUpgrade.pm (it had better not run in the buildfarm client for 15~ and I am wondering if it would be worth backpatching the TAP test once it brews a bit more). Anyway, seeing my time sheet for the next couple of days coupled with a potential beta2 in the very short term and with the broken upgrade workflow, I have given priority to fix the issue because that's what impacts directly people looking at 15 and testing their upgrades, which is what Tushar did. Saying that, I have already sent a pull request to the buildfarm repo to refresh the set of logs, as of the patch attached. This updates the logic so as this would work for any changes in the structure of pg_upgrade_output.d/, fetching any files prefixed by ".log". -- Michael
Commits
Same data as JSON:
GET /api/v1/messages/:b64id/commits
the thread's linked commits as JSON, with link sources.
API reference →
-
Restructure pg_upgrade output directories for better idempotence
- 4fff78f00910 15.0 landed