prefork MPM: fix "httpd -k restart" leaving the server unable to serve - #727
Open
notroj wants to merge 4 commits into
Open
prefork MPM: fix "httpd -k restart" leaving the server unable to serve#727notroj wants to merge 4 commits into
notroj wants to merge 4 commits into
Conversation
a bug which never reached a release needs no entry. No functional change. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
killing the previous generation, which is signalled by process group and so includes the parent itself. Fixes: 243c5fa (SVN r1892587) Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Moved here from the mod_systemd tests, for use by other suites. No functional change. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
An ungraceful restart under prefork leaves the server with no children.
The listening socket stays open, so connections are accepted and never
answered: requests hang rather than fail.
prefork_run() kills the previous generation with
ap_unixd_killpg(getpgrp(), SIGHUP). The parent is in that group. Its
handler, sig_restart() in os/unix/unixd.c, has a guard against taking a
signal it is already handling, but unset_signals() has cleared
restart_pending by then, as a pconf cleanup during the config reload. So
the parent takes its own signal as a fresh restart request and loops:
kill children, come up, restart, repeat. The file comment at prefork.c:146
has always said the root process ignores this signal; it doesn't.
Introduced by 243c5fa (SVN r1892587), which moved the stop of the old
generation to after the config reload, and so to the far side of the flag
being cleared. Reproducible on trunk with a three-module config and no
mod_systemd involved:
leaves zero children and a request that times out. 2.4.x is not affected:
it still stops children before the reload, where the guard does its job.
Also here:
restart with the parent pid unchanged, and two restarts in a row. It is
MPM-generic, so it runs under whichever MPM the job builds. Verified
under both event and prefork.
tests into HttpdTestEnv, for the above.
and a bug which never reached a release needs no entry. Hence no changes
entry here, this regression being trunk-only.
Not addressed: motorz has the identical killpg pattern at motorz.c:2417
and is likely affected, but it is experimental and I have not built or
tested it. Separately, restarting twice quickly makes mod_cgid fail to
bind its socket (AH01243) because the new daemon starts before the old
has unlinked it; the new test ignores that message rather than fixing it.
🤖 Generated with Claude Code