Errors on new VPS

**jagged** · 10-15-2007, 06:21 AM

Hi,
I got an Enterprise VPS deal a few weeks ago and I am trying to fault find an issue with very slow performance when the main domain has hosting turned on. I've got nowhere with support so I thought I'd ask here to see if anyone can shed light on my problem (before I move the sites back to my old Freedom VPS which worked fine).

From my many error logs:

[warn] RSA server certificate CommonName (CN) `plesk' does NOT match server name!?

[warn] Init: SSL server IP/port conflict: default-##-##-###-##:443 (/etc/httpd/conf.d/zz010_psa_httpd.conf:83) vs. webmail:443 (/etc/httpd/conf.d/zz010_psa_httpd.conf:184)
[Mon Sep 24 08:43:26 2007]

[warn] Init: You should not use name-based virtual hosts in conjunction with SSL!!

[Thu Sep 27 11:31:23 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:31:48 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:32:03 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:34:03 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:35:16 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:36:01 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:36:28 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:37:33 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:37:34 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:37:35 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:38:15 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 11:38:42 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 12:02:39 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Thu Sep 27 12:02:40 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/..............................ETC

[Mon Oct 15 04:47:44 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Mon Oct 15 04:47:45 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Mon Oct 15 04:48:10 2007] [error] server reached MaxClients setting, consider raising the MaxClients setting
[Mon Oct 15 04:48:26 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/
[Mon Oct 15 04:48:28 2007] [error] [client 127.0.0.1] Directory index forbidden by Options directive: /var/www/html/..............................ETC

Thanks in advance!
Jagged

**jagged** · 10-15-2007, 06:39 AM

And from my server logs:

127.0.0.1 - - [15/Oct/2007:05:00:26 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:00:27 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:00:28 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:00:29 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:06:08 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:06:09 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:06:11 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:06:12 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:06:14 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:18:16 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:18:17 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:18:18 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:26:21 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:26:22 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:27:12 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:27:14 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"
127.0.0.1 - - [15/Oct/2007:05:27:15 -0400] "GET / HTTP/1.0" 403 5044 "-" "Apache/2.2.3 (CentOS) (internal dummy connection)"...................ETC

**thisisit3** · 10-15-2007, 06:41 AM

You are not giving us enough information to help you.

What exactly do you mean by "very slow performance"? which service is running slow? or is the entire VPS slowing down?

The above log shows that you are running plesk, but i can't say if those errors are related to performance loss.

**jagged** · 10-15-2007, 07:21 AM

Hi,
Sorry about that; I posted the fragments of logs to see if they signalled any obvious issues?

My problem is that for a week or two the new VPS seemed fine (after some initial problems) but at the end of last week I started getting reports from visitors to my 5 hosted sites that page loading was very slow (45 seconds+). I contacted Jaguar support who suggested I needed more memory: "I have checked that your server is lakes in memory. The parameter "provvmpages" is hitting on your server. In order to fix this issue you need extra RAM"

I wasn't convinced by this as the same 5 sites have been working fine on my old Freedom plan with a 1/3 of the memory (I only moved as I thought it would be a faster server) and traffic has been slightly lower than normal. I found that if I turn off the main hosting account the other 4 sites instantly speed up to sub 2 second page loading (even a Drupal based one!). Turning it back on instantly slows them down to a crawl again. I checked and double checked the source code (which was working fine last week), stripped out any unnecessary includes etc and it seemed to be working fine again. However, an hour later all the sites where back to a crawl. I contacted Jaguar support but by the time they checked the server it was running fine again and they suggested it was a network error at my end, even though I could still access my main site on the old Freedom VPS and I haven't been having any issues logging in to Plesk

I'm pretty sure it is a server configuration problem but I don't want to start making changes if they mask another problem and Jaguars last two responses have returned to the 'it's working fine at our end' variety.

Poking around the server settings do you think the Maxclient/server settings could be part of the problem? They seem awfully low but it's difficult to find definitive advice on what the safe limits are?

Code:

# prefork MPM
# StartServers: number of server processes to start
# MinSpareServers: minimum number of server processes which are kept spare
# MaxSpareServers: maximum number of server processes which are kept spare
# ServerLimit: maximum value for MaxClients for the lifetime of the server
# MaxClients: maximum number of server processes allowed to start
# MaxRequestsPerChild: maximum number of requests a server process serves
<IfModule prefork.c>
StartServers       1
MinSpareServers    1
MaxSpareServers    5
ServerLimit       10
MaxClients        10
MaxRequestsPerChild  4000
</IfModule>

# worker MPM
# StartServers: initial number of server processes to start
# MaxClients: maximum number of simultaneous client connections
# MinSpareThreads: minimum number of worker threads which are kept spare
# MaxSpareThreads: maximum number of worker threads which are kept spare
# ThreadsPerChild: constant number of worker threads in each server process
# MaxRequestsPerChild: maximum number of requests a server process serves
<IfModule worker.c>
StartServers       1
MaxClients        10
MinSpareThreads    1
MaxSpareThreads    4
ThreadsPerChild     25
MaxRequestsPerChild  0
</IfModule>

Cheers
Jagged

**Frank Broughton** · 10-15-2007, 08:43 AM

Have you compared the settings above with the ones on your freedom account?

**jagged** · 10-15-2007, 09:52 AM

Originally Posted by Frank Broughton

Have you compared the settings above with the ones on your freedom account?

Hi,
I did that as soon as I noticed (and posted) the low figures and there are lots of differences in the httpd.conf, not just in the MPMs. I can't believe these are the defaults for the Enterprise VPS? There was some confusion when I first got the account as it wasn't the uprated Enterprise plan I paid for and support 'applied the enterprise package specs' to the VPS after I started using it. I'm wondering if I've ended up with some sort of bastard VPS?

Here are the settings in my Freedom (256mb) account:

Code:

# prefork MPM
# StartServers: number of server processes to start
# MinSpareServers: minimum number of server processes which are kept spare
# MaxSpareServers: maximum number of server processes which are kept spare
# ServerLimit: maximum value for MaxClients for the lifetime of the server
# MaxClients: maximum number of server processes allowed to start
# MaxRequestsPerChild: maximum number of requests a server process serves
<IfModule prefork.c>
StartServers       5
MaxClients        50
MinSpareServers    5
MaxSpareServers    10
ServerLimit       50
MaxRequestsPerChild  4000
</IfModule>

# worker MPM
# StartServers: initial number of server processes to start
# MaxClients: maximum number of simultaneous client connections
# MinSpareThreads: minimum number of worker threads which are kept spare
# MaxSpareThreads: maximum number of worker threads which are kept spare
# ThreadsPerChild: constant number of worker threads in each server process
# MaxRequestsPerChild: maximum number of requests a server process serves
<IfModule worker.c>
StartServers         5
MaxClients          50
MinSpareThreads      25
MaxSpareThreads      50 
ThreadsPerChild     25
MaxRequestsPerChild  0
</IfModule>

I'm going to back up the file and do some more research before making changes. When the server is fast it is noticeably faster than my old VPS and the Plesk CP absolutely flys, it's just a shame it's so flaky.

It would really help if someone else with an Enterprise package could share some of their settings? For example, is my Memory allocation limit correct?
privvmpages / Soft Limit 262,144 / Hard Limit 292,912 / Units 4KB pages

Cheers
Jagged

**thisisit3** · 10-15-2007, 12:22 PM

About your apache limits, they are TOO LOW! its definitely a problem and processes will start dying, you don't need more memory, you need to raise those limits.

For example (medium sized webserver):
Timeout 180
KeepAlive On
MaxKeepAliveRequests 100
KeepAliveTimeout 5
MinSpareServers 5
MaxSpareServers 25
StartServers 5
MaxClients 150
MaxRequestsPerChild 8000

Thats for Apache 1.x, so you'd need to convert those values to Apache 2.x.

If you are interested to find out the correct memory/cpu limits for apache, here is a nice script (but you shouldn't need them, unless your sites are VERY busy):

Apache memory limit (RlimitMEM)

**jagged** · 10-16-2007, 04:37 AM

Originally Posted by thisisit3

If you are interested to find out the correct memory/cpu limits for apache, here is a nice script (but you shouldn't need them, unless your sites are VERY busy):

Apache memory limit (RlimitMEM)

Thanks for the advice.

I'm planning on moving a fairly busy site to the VPS once it is running smoothly so I thought I'd give the script a go, but I get errors trying to run even simple perl scripts - something else I need to spend an afternoon checking out...

Cheers
Jagged

Code:

Can't locate SafeFile.pm in @INC (@INC contains: /scripts /usr/lib/perl5/site_per
rl/5.8.8/i386-linux-thread-multi /usr/lib/perl5/site_perl/5.8.7/i386-linux-thread
d-multi /usr/lib/perl5/site_perl/5.8.6/i386-linux-thread-multi /usr/lib/perl5/sit
te_perl/5.8.5/i386-linux-thread-multi /usr/lib/perl5/site_perl/5.8.8 /usr/lib/per
rl5/site_perl/5.8.7 /usr/lib/perl5/site_perl/5.8.6 /usr/lib/perl5/site_perl/5.8.5
5 /usr/lib/perl5/site_perl /usr/lib/perl5/vendor_perl/5.8.8/i386-linux-thread-mul
lti /usr/lib/perl5/vendor_perl/5.8.7/i386-linux-thread-multi /usr/lib/perl5/vendo
or_perl/5.8.6/i386-linux-thread-multi /usr/lib/perl5/vendor_perl/5.8.5/i386-linux
x-thread-multi /usr/lib/perl5/vendor_perl/5.8.8 /usr/lib/perl5/vendor_perl/5.8.7 
 /usr/lib/perl5/vendor_perl/5.8.6 /usr/lib/perl5/vendor_perl/5.8.5 /usr/lib/perl5
5/vendor_perl /usr/lib/perl5/5.8.8/i386-linux-thread-multi /usr/lib/perl5/5.8.8 .
.) at ./file.pl line 3.B
BEGIN failed--compilation aborted at ./file.pl line 3.[

**jagged** · 10-16-2007, 01:06 PM

Well I made the changes to the config but the server still keeps slowing to a crawl

I checked using Sitespeed on iwebtool.com and here are the results for the front pages on two of my sites, which use the same news scripts but one has been stripped right back to try and speed it up. Take into account the site on the freedom VPS gets around 10X the traffic!

1 xxx.com 22.18 KB 10.19 seconds - enterprise VPS

2 xxx.com 39.69 KB 0.73 seconds - old freedom VPS

I'll give support another chance to look at this but if they keep saying its OK then I'm going to have to migrate back to the Freedom VPS for now and I'll look elsewhere for a faster server

Cheers
Jagged

PS I just checked again and I got 0.5 second load times followed by 8 seconds then 45!!!

**thisisit3** · 10-17-2007, 01:13 AM

Ask someone from the senior tech support guys, like JPC-Masood, because the others won't be of any help to you.

Either the VPS node has a problem (there was a kernel with that kind of problems but it should have been replaced), or your software is causing the problem due to a misconfiguration.

**jagged** · 10-17-2007, 03:22 AM

I don't think I'll bother contacting support about this issue unless I can pinpoint the problem. They appear to be annoyed at me wasting their time.

The latest response was 'The load on both Hardware node and on your VPS is normal and I have confirmed there is no network issue.....Since the issue is not at our side, we can't troubleshoot it further'. So I guess I don't need more memory after all?

The problem isn't at my end either as 6 different web site speed tests where reporting 30s + loading times for a 20Kb page, so they suggested 'There is a possibility that the problem lies somewhere in the middle' and I should 'wait for some time and hopefully the issue will be resolved'.

At least those damn middle men haven't been messing with my old Freedom VPS

Cheers
Jagged

**JPC-Veena** · 10-17-2007, 05:20 AM

Jagged,

Sorry for the problems you are seeing.
Please provide here the ticket number so we can take a look at it. Also don't forget to update the ticket with the link to the main domain which has problems and what software you are running on that website in case it doesn't have the info already.

**Ron** · 10-17-2007, 09:06 AM

And run winMTR to prove/disprove the man in the middle problem...

**jagged** · 10-19-2007, 01:32 AM

Well I think I've fixed the problem. I painstakingly went through every option in the httpd.conf and php.ini checking them against my Freedom VPS standard configuration and various online resources and the VPS seems to work fine now

Unfortunately I awoke this morning to find that my old Freedom VPS is down - no www/ftp/ssh/plesk etc and site uptime's servers report the same from across the pond. Back on the support wagon again

Cheers
Jagged

**jagged** · 10-19-2007, 04:02 AM

Apparently my Freedom VPS has been down for the last four hours because of backups causing a heavily loaded node. I guess I'll have to go and do some DIY or something while I wait for it to come back online

Cheers
Jagged

LinkBack

Thread Tools

Display

Errors on new VPS

Bookmarks

Bookmarks

Posting Permissions