tracker issue : CF-4205438

select a category, or use search below
(searches all categories and all time range)
Title:

Update 5 Connector @ CentOS 7 and Apache Configuration Problems

| View in Tracker

Status/Resolution/Reason: To Track//NeedMoreInfo

Reporter/Name(from Bugbase): MLADEN C. / ()

Created: 10/15/2019

Components: connector

Versions: 2018

Failure Type:

Found In Build/Fixed In Build: 2018,0,05,315699 /

Priority/Frequency: Normal / Unknown

Locale/System: English / Linux

Vote Count: 0

Problem Description:

Connector recreation is required after updating CF 2018 to Update 5. That causes the Apache to randomly fails to start.

Steps to Reproduce:

Install CentOS 7, Apache, ColdFusion 2018, Update 4 (required before Update 5), Update 5, recreate connector (extract mod_jk.so from {cf-installation-dir}/cfusion/runtime/lib/wsconfig.jar and copy to required dir eg. /etc/httpd/conf), restart CF services (or entire server) and then restart/check status of Apache (httpd). If it's failed go and check journalctl -xe or systemctl status httpd. Command dmesg also shows httpd error. If it's started try to restart it few times with systemctl restart httpd. It will fail.

Actual Result:
Randomly failing Apache (httpd) service

Expected Result:
Working Apache (httpd) service, starte/stopped/restarted without any problems any time

Any Workarounds:
Using connector from older updates (Update 4 OK)
Connector version 1.2.46.314600 that comes with Update 5 has size (in linux) 1266464 bytes and causes problems with Apache (in my build).
Connector version 1.2.46.314600 that is recommended by Adobe if people want to use remote port to access ColdFusion Admin and is for Update 5 also (can be downloaded from https://documentcloud.adobe.com/link/track?uri=urn%3Aaaid%3Ascds%3AUS%3Ae7ba3112-ebf0-4190-b2ea-b786c1bf73ea) has size 1266381 bytes also causes problems with Apache (in my build).
Connector version 1.2.46.312584 and comes with Update 4 has size 1259080 and works perfectly fine.
Checking the minor version (312584, 314600) of the connector (mod_jk.so 1.2.46) can be done by reading /etc/httpd/conf/mod_jk_global.log log file.

Attachments:

Comments:

MLADEN C., can you pls share error msg that you see when HTTPD fails to start. Just to be clear, the problem is intermittent, when you restart HTTPD, it starts sometimes but fails to start at other time. And when it is started it has no issues servicing CFM requests. Also what's the version and build of the Apache HTTPD you are using.
Comment by Piyush K.
31635 | October 17, 2019 02:34:59 PM GMT
Hi Piyush, I am very busy with projects at work. I had to abandon the CF2018 test server and currently testing CF2016. As soon as I grab more time I will recreate CF2018 and post the errors you asked for. The version of httpd is latest for the CentOS 7 - 2.4.6-90.el7.centos Out of let's say 10 tries to start the httpd, one of those is successful and then there's no problems with CFM requests. However if I restart my server I must start httpd manually (from many tries).
Comment by MLADEN C.
31646 | October 18, 2019 01:32:12 PM GMT
One of the errors can be seen using 'dmesg' command httpd[26283]: segfault at 7f87928a7220 ip 00007f879e95738a sp 00007f87926309c8 error 4 in libc-2.17.so[7f879e8c1000+1c3000] ? httpd.service - The Apache HTTP Server Loaded: loaded (/usr/lib/systemd/system/httpd.service; enabled; vendor preset: disabled) Active: failed (Result: exit-code) since Fri 2019-10-18 12:18:19 EDT; 2s ago Docs: man:httpd(8) man:apachectl(8) Process: 26335 ExecStop=/bin/kill -WINCH ${MAINPID} (code=exited, status=1/FAILURE) Process: 26332 ExecStart=/usr/sbin/httpd $OPTIONS -DFOREGROUND (code=killed, signal=SEGV) Main PID: 26332 (code=killed, signal=SEGV) Oct 18 12:18:19 systemd[1]: Starting The Apache HTTP Server... Oct 18 12:18:19 systemd[1]: httpd.service: main process exited, code=killed, status=11/SEGV Oct 18 12:18:19 kill[26335]: kill: cannot find process "" Oct 18 12:18:19 systemd[1]: httpd.service: control process exited, code=exited status=1 Oct 18 12:18:19 systemd[1]: Failed to start The Apache HTTP Server. Oct 18 12:18:19 systemd[1]: Unit httpd.service entered failed state. Oct 18 12:18:19 systemd[1]: httpd.service failed.
Comment by MLADEN C.
31647 | October 18, 2019 04:20:00 PM GMT
Thanks for sharing the error. If you comment out the directive that includes mod_jk.conf (one of the last lines in httpd.conf), does HTTPD start without any errors, every time? ps. restarting CF server does not restart your web-server.
Comment by Piyush K.
31653 | October 19, 2019 07:28:36 AM GMT
Piyush, If I comment that out, it restarts. If you read it from start you'll see that other mod_jk.so files are working. Recently I have noticed the same behavior with ColdFusion 2016 and 314600 version of mod_jk.so too, but less often, while 313469 (version that allows you to use remote ports works perfectly). I thought I was clear enough in my bug post. However I know that restarting CF does not restart httpd (Apache) otherwise I would not write here Also I meant if restarting whole VM (server, machine, box - name it), is not reliable with 314600 version of mod_jk.so because Apache may not start each time. That I consider as unreliable server then.
Comment by MLADEN C.
31666 | October 21, 2019 03:50:52 PM GMT