> Could you please look if the test can be made more resilient against such > slowdowns?
Attached a single line change that resets the timer for each query for the background connection. I don't see anything wrong with the tests otherwise, the problem is simply that we are using a single background connection for a long time, and on a heavily overloaded machine with asan, the test can sometimes exceed the timeout value.
From eee33c5e3b385a2126048643ad4f05e770d39b86 Mon Sep 17 00:00:00 2001 From: Zsolt Parragi <[email protected]> Date: Sun, 8 Nov 2026 10:12:06 +0000 Subject: [PATCH v1] oauth: Apply bgconn timeout per query in 001_server The background psql session in oauth_validator/001_server lives for most of the test, but its timer was started once and never restarted, so timeout_default applied to the whole session rather than to each query. On slow or loaded buildfarm animals the test can exceed that and the next query fails with "IPC::Run: timeout on timer #1" even though nothing is stuck. Fix by restarting the timer per query. Backpatch to v18 where the test was introduced. Reported-by: Alexander Lakhin <[email protected]> Discussion: https://postgr.es/m/[email protected] Backpatch-through: 18 --- src/test/modules/oauth_validator/t/001_server.pl | 1 + 1 file changed, 1 insertion(+) diff --git a/src/test/modules/oauth_validator/t/001_server.pl b/src/test/modules/oauth_validator/t/001_server.pl index 5308550d06c..69c58cec706 100644 --- a/src/test/modules/oauth_validator/t/001_server.pl +++ b/src/test/modules/oauth_validator/t/001_server.pl @@ -60,6 +60,7 @@ my $exec_backend = $node->safe_psql('postgres', 'SHOW debug_exec_backend'); # Save a background connection for later configuration changes. my $bgconn = $node->background_psql('postgres'); +$bgconn->set_query_timer_restart(); my $webserver = OAuth::Server->new(); $webserver->run(); -- 2.56.0
