https - 443 stop working often

[jeroman8]

During the past 2 weeks I have had issues with https.
It just stop working suddenly and you need to restart httpd to get it up again.
It's only https that is down, http is working great.

I tested to go to: http://www.domain.com:443 (without ssl - https) but that
is not working.

On a server where https is running you can access :443, there's a message there
but it's like the port is closed when https is not working.

Why and how can a port suddenly be closed and is this the cause of why https
is not working or is the port closed because https stopped ?

I checked the ssl logs and fist there was a conflict, I had two copies of a ssl domain
in httpd config so I fixed this but it still keeps crashing.
APF firewall has 443 open and even if I stop APF https will not work again if it
once was down so I do not think this is a firewall issue.

Any ideas ?

 

[rhenderson]

No Ideas but just started to encounter the same problem, looking forward to anyones response.

 

[chirpy]

One possible reason - If you're restarting httpd manually, make sure you use:

httpd startssl

and not

httpd start

If it's a port block issue, you can check exactly which daemon has it open with:

netstat -lpn

It should obviously be httpd.

Last, but not least, make sure that you only have 1 virtualhost container using the IP address for each SSL site.

 

[rhenderson]

One possible reason - If you're restarting httpd manually, make sure you use:

httpd startssl

and not

httpd start

If it's a port block issue, you can check exactly which daemon has it open with:

netstat -lpn

It should obviously be httpd.

Last, but not least, make sure that you only have 1 virtualhost container using the IP address for each SSL site.

Checked the IP for each SSL site. With the httpd startssl, does that work much differently than the service httpd restart or /scripts/restartsrv_httpd ?? Just curious.

My problems started when I went to kill the cache on eaccelerator and killed the whole directory then httpd would not restart, check the logs saw the problem remade the directory, gave it permissions and used service httpd restart, it complained it was not running then started apache. When it was not working, port 443 was not open at all (checked with netstat), first checked to make sure nothing changed in the conf.apf then did a httpd stop then httpd start, then port 443 came up, double checked with netstat.

I don't ever remember having to use the httpd startssl. I am assuming the httpd startssl starts apache? or just the ssl part?

Thanks for all your help Chirpy.

 

[jeroman8]

I always use /scripts/restartsrv_httpd when I do it manually.
SSL works fine then so I guess that script takes care of ssl also.