Bug #18316: Rerun job with failed hosts fails with "Stack level too deep" in the log. - Foreman Remote Execution - Foreman

Bug #18316

Rerun job with failed hosts fails with "Stack level too deep" in the log.

Added by Shimon Shtein over 1 year ago. Updated 9 days ago.

Status:	Closed
Priority:	Normal
Assignee:	Shimon Shtein
Category:	-
Target version:	foreman_remote_execution 1.3.1
Difficulty:		Team Backlog:
Triaged:		Fixed in Releases:
Bugzilla link:	1406489	Found in Releases:
Pull request:	https://github.com/theforeman/foreman_remote_execution/pull/226

Description

Started as a reproducer for https://bugzilla.redhat.com/show_bug.cgi?id=1406489
The test didn't reproduce the issue, but it's already written.

Associated revisions

Revision 68719465
Added by Shimon Shtein over 1 year ago

Fixes #18316 - Changed scoped_search query to :in for hosts enum

When asking to target a specific set of hosts (like in case of failed
hosts), a `name = X or name = Y...` query was generated. This query
caused a very deep stack recursion in scoped_search AST. This commit
changes the query to `name ^ (X, Y...)` that will not require the same
treatment in AST.

History

#1 Updated by The Foreman Bot over 1 year ago

Status changed from New to Ready For Testing
Pull request https://github.com/theforeman/foreman_remote_execution/pull/226 added

#2 Updated by Ivan Necas over 1 year ago

Target version set to 1.11.3
Bugzilla link set to 1406489

#3 Updated by Shimon Shtein over 1 year ago

After an investigation, I was able to reproduce the issue. The problem was in query syntax: originally rerun query was generating a huge OR statement like: "name = X or name = Y...". This query could not be loaded into AST inside scoped_search without blowing up the stack.