XFileSharing Pro - need urgent help with apache serverload & index_dl.cgi

Message
Author
stamos
Posts: 139
Joined: Nov 11, 2010 5:37 pm

need urgent help with apache serverload & index_dl.cgi

#1 Postby stamos » Mar 30, 2012 4:50 pm

Since support is not answering maybe someone can help me with this shit serverload.
currently my webserver sever is reaching loads of 200 and more, though I dont have so much traffic.
apache is on a dual quadcore with SSD harddisk, mysql is on it's on server.
any ideas what could help?

what is this shit index_dl.cgi script doing anyway and causes so high load?

hostlife
Posts: 192
Joined: Aug 13, 2011 12:34 pm

#2 Postby hostlife » Mar 30, 2012 6:05 pm

You should get nginx mod. It will reduce your load dramatically.

stamos
Posts: 139
Joined: Nov 11, 2010 5:37 pm

#3 Postby stamos » Mar 30, 2012 6:15 pm

too many problems with it? the forum is full of issues with nginx.
also, it has to go on every fileserver?

chrda
Posts: 296
Joined: Sep 14, 2009 7:16 pm

#4 Postby chrda » Mar 30, 2012 7:04 pm

Nginx works like clockwork.
If there is any problems, they are usually related to tuning or hardware resources.

There is many ways to reduce load, but it demands either a good server admin with some code skills , and/or a perl programmer.

Reason why is the load is high on index_dl.cgi is that the users want all those functions for a fully script. You can tune the script with a programmer after your own needs.

Apache need to be tuned
MySQL need to be tuned
Nginx need to be tuned on high traffic sites.

stamos
Posts: 139
Joined: Nov 11, 2010 5:37 pm

#5 Postby stamos » Mar 30, 2012 8:16 pm

i tuned apache and mysql pretty well,
what I am wondering is why apache get's high loads, 90% cpu use at a dual qudcore(!), and not mysql.
usually in scripts like this mysql gets high loads.
the sql server runs under 1.0 load and the apache server reaches loads up to 200.
and when yesterday the load was 3-4 today it goes up to 200, this is not reasonable is it?!

chrda
Posts: 296
Joined: Sep 14, 2009 7:16 pm

#6 Postby chrda » Mar 30, 2012 8:23 pm

You can try litespeed.

Its very easy to ddos a XFS site, making the cpu go skyhigh :)

You can install CSF firewall and setup the connection tracker also.
Maybe you got some bogus traffic.

glumbo
Posts: 211
Joined: Mar 11, 2011 6:26 am

#7 Postby glumbo » Mar 31, 2012 10:21 am

Email me if you want the fix. It will save you a lot of money on server fees as you won't need a more powerful server. I had a lot of problems with index_dl.cgi as well but I know how to reduce the load by 80-90%

stamos
Posts: 139
Joined: Nov 11, 2010 5:37 pm

#8 Postby stamos » Mar 31, 2012 7:40 pm

thanks glumbo, just send you an email.

though I think besides ddos which you can recognize pretty easy the script itself creates too much load.

cisco firewall is ordered and will be installed from monday on :)

krazymax
Posts: 62
Joined: Jan 31, 2012 8:45 am

#9 Postby krazymax » Apr 13, 2012 7:50 am

am facing a similar issue. i get mail every few minutes
lfd on sr.xxx.com: Suspicious process running under user admin

Time: Fri Apr 13 11:25:31 2012 +0400
PID: 7991
Account: admin
Uptime: 115 seconds

Executable:
/usr/bin/perl

Command Line (often faked in exploits):
/usr/bin/perl index_dl.cgi
my main site is on a vps and fileserver is a dedicated. Traffic is low, so i thought vps would be enough. now its going down every few minutes.

any help"?

stamos
Posts: 139
Joined: Nov 11, 2010 5:37 pm

#10 Postby stamos » Apr 18, 2012 7:39 pm

seems there is no solution.
installing nginx is no option for me due to plesk on the server.
glumbo never responded to my email as well.

I guess just a powerfull server(s) & loadbalancing helps.