Mercurial > hg > nginx-tests
view proxy_next_upstream.t @ 1752:ba6e24e38f03
Tests: improved stop_daemons() to send signal again.
As was observed, it's possible that a signal to complete a uwsgi daemon
can be ignored while it is starting up, which results in tests hang due
to eternal waiting on child processes termination. Notably, it is seen
when running tests with a high number of prove jobs on a low-profile VM
against nginx with broken modules and/or configuration. To reproduce:
$ TEST_NGINX_GLOBALS=ERROR prove -j16 uwsgi*.t
Inspecting uwsgi under ktrace on FreeBSD confirms that a SIGTERM signal
is ignored at the very beginning of uwsgi startup. It is then replaced
with a default action after listen(), thus waiting until uwsgi is ready
to accept new TCP connections doesn't completely solve the hang window.
The fix is to retry sending a signal some time after waitpid(WNOHANG)
continuously demonstrated no progress with reaping a signaled process.
It is modelled after f13ead27f89c that improved stop() for nginx.
author | Sergey Kandaurov <pluknet@nginx.com> |
---|---|
date | Wed, 29 Dec 2021 22:29:23 +0300 |
parents | 5ac6efbe5552 |
children |
line wrap: on
line source
#!/usr/bin/perl # (C) Maxim Dounin # Tests for http proxy module, proxy_next_upstream directive. ############################################################################### use warnings; use strict; use Test::More; BEGIN { use FindBin; chdir($FindBin::Bin); } use lib 'lib'; use Test::Nginx; ############################################################################### select STDERR; $| = 1; select STDOUT; $| = 1; my $t = Test::Nginx->new()->has(qw/http proxy rewrite/)->plan(8); $t->write_file_expand('nginx.conf', <<'EOF'); %%TEST_GLOBALS%% daemon off; events { } http { %%TEST_GLOBALS_HTTP%% upstream u { server 127.0.0.1:8081; server 127.0.0.1:8082; } upstream u2 { server 127.0.0.1:8081; server 127.0.0.1:8082; } upstream u3 { server 127.0.0.1:8081; server 127.0.0.1:8082 down; } server { listen 127.0.0.1:8080; server_name localhost; location / { proxy_pass http://u; proxy_next_upstream http_500 http_404; } location /all/ { proxy_pass http://u2; proxy_next_upstream http_500 http_404; error_page 404 /all/404; proxy_intercept_errors on; } location /all/404 { return 200 "$upstream_addr\n"; } location /down { proxy_pass http://u3; proxy_next_upstream http_404; } } server { listen 127.0.0.1:8081; server_name localhost; location / { return 404; } location /ok { return 200 "AND-THIS\n"; } location /500 { return 500; } location /all/ { return 404; } } server { listen 127.0.0.1:8082; server_name localhost; location / { return 200 "TEST-OK-IF-YOU-SEE-THIS\n"; } location /all/ { return 404; } } } EOF $t->run(); ############################################################################### my ($p1, $p2) = (port(8081), port(8082)); # check if both request fallback to a backend # which returns valid response like(http_get('/'), qr/SEE-THIS/, 'proxy request'); like(http_get('/'), qr/SEE-THIS/, 'second request'); # make sure backend isn't switched off after # proxy_next_upstream http_404 like(http_get('/ok') . http_get('/ok'), qr/AND-THIS/, 'not down'); # next upstream on http_500 like(http_get('/500'), qr/SEE-THIS/, 'request 500'); like(http_get('/500'), qr/SEE-THIS/, 'request 500 second'); # make sure backend switched off with http_500 unlike(http_get('/ok') . http_get('/ok'), qr/AND-THIS/, 'down after 500'); # make sure all backends are tried once like(http_get('/all/rr'), qr/^127.0.0.1:($p1, 127.0.0.1:$p2|$p2, 127.0.0.1:$p1)$/mi, 'all tried once'); # make sure backend marked as down doesn't count towards "no live upstreams" # after all backends are tried with http_404 like(http_get('/down/'), qr/Not Found/, 'all tried with down'); ###############################################################################