Closed
Bug 1272480
Opened 9 years ago
Closed 9 years ago
OrangeFactor bug commenter broken since nginx changes on 2016-05-03
Categories
(Tree Management Graveyard :: OrangeFactor, defect)
Tree Management Graveyard
OrangeFactor
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: jwalker, Assigned: emorley)
References
Details
This link https://brasstacks.mozilla.com/orangefactor/?display=OrangeFactor&includefiltertype=quicksearch&includefilterdetailsexcludeResolved=false&includefilterdetailsexcludeDisabled=false&includefilterdetailsquicksearch=&includefilterdetailsnumbugs=0&includefilterdetailsresolvedIds=&excludefiltertype=quicksearch&excludefilterdetailsquicksearch=&excludefilterdetailsnumbugs=0&excludefilterdetailsresolvedIds=&plat=All&test=mochitest-devtools-chrome-8&type=All&startday=2016-05-05&endday=2016-05-12&tree=trunk
Currently gives me the following:
bug 1131012 count=35 (can believe this one)
bug 1265565 count=16 (single occurrence according to Bugzilla)
bug 1269977 count=15 (single occurrence according to Bugzilla)
Comment 2•9 years ago
|
||
The data that appears in bugzilla is aggregated periodically (weekly?) and so it's expected that the most recent occurrences may not appear there.
It also only comments if you reach a daily or weekly minimum threshold of failures.
Assignee | ||
Comment 4•9 years ago
|
||
The thresholds and periodic nature are true (for more details see: ), however in this instance the bug commenter has actually started failing. The last successful run was at 2016-05-03 02:00 UTC+1..
Getting bug data...
-> Fetching JSON from http://localhost/orangefactor/api/bybug?tree=all&startday=2016-05-02&endday=2016-05-02
Submitting comment to bug 1268163 (15 occurrences)
...
Complete!
Getting bug data...
-> Fetching JSON from http://localhost/orangefactor/api/bybug?tree=all&startday=2016-05-03&endday=2016-05-03
HTTPError 404 fetching http://localhost/orangefactor/api/bybug?tree=all&startday=2016-05-03&endday=2016-05-03: <html>
<head><title>404 Not Found</title></head>
<body bgcolor="white">
<center><h1>404 Not Found</h1></center>
<hr><center>nginx/1.2.8</center>
</body>
</html>
...which was just prior to :atoll changing the nginx config (and recompiling nginx) to fix the certificate issues with the custom version of nginx that was running on there previously:
428 2016-04-28 19:06:26 vmware-config-tools.pl
429 2016-05-03 17:13:31 cd /etc/nginx/
430 2016-05-03 17:13:32 ls -al
431 2016-05-03 17:13:33 cd conf.d
432 2016-05-03 17:13:34 ls -al
433 2016-05-03 17:13:35 vim orangefactor.conf
434 2016-05-03 17:13:43 cd /usr/local/nginx/ssl
435 2016-05-03 17:13:43 ls -al
436 2016-05-03 17:13:46 less brasstacks.mozilla.org.pem
437 2016-05-03 17:13:51 cp brasstacks.mozilla.org.pem ~/
438 2016-05-03 17:13:52 cd ~/
439 2016-05-03 17:13:53 vim brasstacks.mozilla.org.pem
440 2016-05-03 17:14:01 cd -
441 2016-05-03 17:14:02 ls -al
442 2016-05-03 17:14:04 mv brasstacks.mozilla.org.pem{,.orig}
443 2016-05-03 17:14:06 mv ~/brasstacks.mozilla.org.pem .
444 2016-05-03 17:14:07 ls -al
445 2016-05-03 17:14:10 service nginx graceful
446 2016-05-03 17:14:14 service nginx restart
447 2016-05-03 17:20:29 ls -al
448 2016-05-03 17:20:31 cd /etc/httpd
449 2016-05-03 17:20:32 ls -al
450 2016-05-03 17:20:33 cd /etc/nginx
451 2016-05-03 17:20:35 ls -al
452 2016-05-03 17:20:40 vim nginx.conf
453 2016-05-03 17:20:44 cd conf.
454 2016-05-03 17:20:46 cd conf.d
455 2016-05-03 17:20:46 ls -al
456 2016-05-03 17:20:47 vim orangefactor.conf
457 2016-05-03 17:21:34 service nginx restart
458 2016-05-03 17:22:34 fg
459 2016-05-03 17:23:57 service nginx restart
460 2016-05-03 17:24:03 fg
461 2016-05-03 17:24:49 service nginx restart
462 2016-05-03 17:25:23 ls -al
463 2016-05-03 17:25:25 fg
464 2016-05-03 17:26:25 cd /var/log/nginx
465 2016-05-03 17:26:26 ls -al
466 2016-05-03 17:26:27 tail error.log
467 2016-05-03 17:26:39 date
468 2016-05-03 17:28:26 dpkg -L | grep openssl
469 2016-05-03 17:35:37 vim orangefactor.conf
470 2016-05-03 17:35:39 cd -
471 2016-05-03 17:35:40 vim orangefactor.conf
472 2016-05-03 17:35:48 service nginx restart
473 2016-05-03 17:35:58 fg
474 2016-05-03 17:38:03 apt-get install openssl
475 2016-05-03 17:38:05 yum
476 2016-05-03 17:38:08 yum install openssl
477 2016-05-03 17:38:57 openssl ec
478 2016-05-03 17:39:07 rpm -qi nginx
479 2016-05-03 17:39:09 yum install nginx
480 2016-05-03 17:53:03 cd ~/
481 2016-05-03 17:53:06 yum upgradeonly
482 2016-05-03 17:53:10 yum install --upgradeony
483 2016-05-03 17:53:12 yum install --upgradeonly
484 2016-05-03 17:53:15 yum install --help
485 2016-05-03 17:53:30 yum upgrade --help
486 2016-05-03 17:53:32 ls
487 2016-05-03 17:53:36 rpm -qa | grep nginx
488 2016-05-03 17:53:38 cd ~rsoderberg/
489 2016-05-03 17:53:39 ls
490 2016-05-03 17:54:22 yum upgrade nginx-1.2.8-1.el6.ngx.x86_64.rpm
491 2016-05-03 17:54:33 service nginx
492 2016-05-03 17:54:35 service nginx configtest
493 2016-05-03 17:54:40 service nginx restart
494 2016-05-03 17:55:13 cd /etc/nginx
495 2016-05-03 17:55:14 ls
496 2016-05-03 17:55:16 cd conf.d
497 2016-05-03 17:55:16 ls
498 2016-05-03 17:55:18 vim orangefactor.conf
499 2016-05-03 17:55:47 service nginx restart
500 2016-05-03 18:25:48 which nginx
501 2016-05-16 12:19:07 history | less
502 2016-05-16 12:19:28 su - webtools
503 2016-05-16 12:30:38 history | tail -n 100
IRC log:
18:09 <atoll> is httpd intentionally off?
18:09 <emorley> we're using nginx
18:09 <atoll> the site is down
18:09 <emorley> wfm? https://brasstacks.mozilla.com/orangefactor/
18:09 <atoll> so.. if that's cool, wfm
18:09 <atoll> huh
18:10 <atoll> whaaat
18:10 — atoll pokes at something
18:10 <atoll> http://brasstacks.mozilla.com
18:10 <atoll> doesn't work for me.
18:10 <atoll> https, either
18:10 <atoll> https://brasstacks.mozilla.com
18:10 <atoll> oh, wow
18:10 <atoll> i bet i have SHA1 disabled or something
18:11 <atoll> nope. trippy.
18:11 <atoll> oh, yeah, this is the one with the damaged SSL config
18:11 <emorley> I have us using the more aggressive config from https://wiki.mozilla.org/Security/Server_Side_TLS#Modern_compatibility
18:11 <atoll> it's apparently incompatible with OS X as deployed, which is interesting
18:12 — atoll pokes around a bit
18:12 <atoll> your intermediate cert is deployed incorrectlyt
18:12 <atoll> let me get a correct file,s ec
18:12 <emorley> thank you
18:13 — atoll general purpose diagnostic machine
18:13 <emorley> someone kindly set it up post-renewal recently
18:14 <emorley> config in /etc/nginx/conf.d/orangefactor.conf
18:14 <atoll> okay, i fixed the intermediate on the server
18:14 <emorley> great :-)
18:14 <atoll> i'll keep looking about the osx thing, but will report back without making changes on that part
18:14 <atoll> more later!
18:15 <emorley> I'll have another look at what SSL labs says
18:15 <atoll> it says A+ now
18:15 <atoll> intermediate is repaired
18:15 <emorley> I'd checked before and it said A+ then, strange
18:16 <atoll> if you look in the output down by IE 11 and Safari, you can see the various handshake_failures, which is interesting to me
18:16 <atoll> yeah, the root in intermediate thing isn't a -
...
18:22 <atoll> i tried removing the single quotes from the cipherstring first, since they might be screwing it up
18:23 <atoll> nope, no change
...
18:25 <atoll> i have ulfr on the other side trying to understand what in our config is breaking this
18:25 <atoll> since he authors the config generator
...
18:37 <atoll> okay, so
18:37 <atoll> your server doesn't support ECDHE
18:37 <atoll> so you aren't capable of running the high security config
18:38 <atoll> and that's why it's collapsing
18:40 <emorley> strange since http://nginx.org/en/CHANGES says nginx 1.1.0 supports ECDHE
18:41 <atoll> yeah, but redhat didn't at the time
18:41 <atoll> and we're running some stupid self compiled nginx
18:41 <atoll> fffff
...
18:52 <atoll> okay, i have nginx 1.2.8 available now.
18:52 <atoll> recompiled with ec support, presumably
18:52 <atoll> shall i found out?
18:52 <atoll> ^find out?
18:52 <emorley> yeah might as well!
18:54 <atoll> there you go
18:54 <atoll> works in safari now
18:55 <atoll> bumping up your ssl config to the latest High one
18:58 <emorley> amazing, thank you :-)
Assignee: nobody → emorley
Component: Treeherder → OrangeFactor
Flags: needinfo?(emorley)
Summary: Filtering OrangeFactor by mochitest-devtools-chrome-8 gives results inconsistent with bugzilla → OrangeFactor bug commenter broken since nginx changes on 2016-05-03
Assignee | ||
Comment 5•9 years ago
|
||
(In reply to Ed Morley [:emorley] from comment #4)
> The thresholds and periodic nature are true (for more details see: )
https://groups.google.com/d/msg/mozilla.dev.tree-management/az643p0u4hs/3el7fqIDBwAJ
--
So the nginx served static files are fine being served from localhost:
[emorley@brasstacks1.dmz.scl3 ~]$ curl -is http://localhost/orangefactor/ | head -n 20
HTTP/1.1 200 OK
Server: nginx/1.2.8
Date: Mon, 16 May 2016 12:44:21 GMT
Content-Type: text/html
Content-Length: 10757
Last-Modified: Tue, 15 Mar 2016 16:28:41 GMT
Connection: keep-alive
Accept-Ranges: bytes
<!DOCTYPE html>
<html>
<!-- This Source Code is subject to the terms of the Mozilla Public License
- version 2.0 (the "License"). You can obtain a copy of the License at
- http://mozilla.org/MPL/2.0/. -->
<head>
<title>OrangeFactor</title>
<link rel="icon" type="image/vnd.microsoft.icon" href="favicon.ico" />
<link rel="stylesheet" href="style/main.css" type="text/css"/>
<link rel="stylesheet" href="vendor/jquery.datepick.css" type="text/css"/>
<link rel="stylesheet" href="vendor/datatables.css" type="text/css"/>
...just not the fastcgi passthrough to the OrangeFactor API:
[emorley@brasstacks1.dmz.scl3 ~]$ curl -is http://localhost/orangefactor/api/bybug | head -n 20
HTTP/1.1 404 Not Found
Server: nginx/1.2.8
Date: Mon, 16 May 2016 12:45:11 GMT
Content-Type: text/html
Content-Length: 168
Connection: keep-alive
<html>
<head><title>404 Not Found</title></head>
<body bgcolor="white">
<center><h1>404 Not Found</h1></center>
<hr><center>nginx/1.2.8</center>
</body>
</html>
The nginx config:
[emorley@brasstacks1.dmz.scl3 ~]$ cat /etc/nginx/conf.d/orangefactor.conf
server {
listen 80;
server_name localhost;
include /etc/nginx/default_locations/*.conf;
}
server {
listen 443;
server_name brasstacks.mozilla.com;
include /etc/nginx/default_locations/*.conf;
ssl on;
ssl_certificate SNIP;
ssl_certificate_key SNIP;
# New config from https://mozilla.github.io/server-side-tls/ssl-config-generator/ ...
ssl_session_timeout 1d;
ssl_session_cache shared:SSL:50m;
# Diffie-Hellman parameter for DHE ciphersuites, recommended 2048 bits
ssl_dhparam SNIP;
# Updated modern. 20160503 atoll
ssl_protocols TLSv1.2;
ssl_ciphers SNIP;
ssl_prefer_server_ciphers on;
# HSTS (ngx_http_headers_module is required) (15768000 seconds = 6 months)
add_header Strict-Transport-Security max-age=15768000;
}
The locations config:
[emorley@brasstacks1.dmz.scl3 ~]$ cat /etc/nginx/default_locations/orangefactor.conf
location / {
root /usr/share/nginx/html;
index index.html index.htm;
rewrite ^/$ https://brasstacks.mozilla.com/orangefactor/ permanent;
rewrite ^/index.html$ https://brasstacks.mozilla.com/orangefactor/ permanent;
}
location /orangefactor/api/ {
fastcgi_ignore_client_abort on;
fastcgi_read_timeout 300;
fastcgi_param REQUEST_METHOD $request_method;
fastcgi_param QUERY_STRING $query_string;
fastcgi_param CONTENT_TYPE $content_type;
fastcgi_param CONTENT_LENGTH $content_length;
fastcgi_param GATEWAY_INTERFACE CGI/1.1;
fastcgi_param SERVER_SOFTWARE nginx/$nginx_version;
fastcgi_param REMOTE_ADDR $remote_addr;
fastcgi_param REMOTE_PORT $remote_port;
fastcgi_param SERVER_ADDR $server_addr;
fastcgi_param SERVER_PORT $server_port;
fastcgi_param SERVER_NAME $server_name;
fastcgi_param SERVER_PROTOCOL $server_protocol;
fastcgi_param SCRIPT_FILENAME $fastcgi_script_name;
fastcgi_split_path_info ^(/orangefactor)(.*)$;
fastcgi_param PATH_INFO $fastcgi_path_info;
fastcgi_pass 127.0.0.1:9500;
}
The root account history doesn't show any changes being made to the locations config, so guessing there must have been a breaking change as part of updating from nginx 1.2.0 to 1.2.8 ? Looking at the changelog there only seem to have been 'fixes' to bugs in the fastcgi handling however:
http://nginx.org/en/CHANGES-1.2
Assignee | ||
Comment 7•9 years ago
|
||
I've just worked around this - and updated woo_cron.conf such that the cron jobs use:
local_server_url = https://brasstacks.mozilla.com/orangefactor/api
...rather than:
local_server_url = http://localhost/orangefactor/api
I've re-run the daily and weekly bug comment jobs and also the weekly email job now (though no easy way to re-run for prior days, but I think that's fine - the original data is still available on OrangeFactor).
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Updated•5 years ago
|
Product: Tree Management → Tree Management Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•