I'm have a VDS with hosting-ie, and it is extremely slow.. I run almost nothing on it..
I'm not running a website, no MYSQL, no Databases.. just newsleecher, a bittorent client for seeding ubunto and occassionally i'll open ftp program to transfer files from home to server.
Lately, i can't even just have newsleecher open by itself and have it running, as it pauses waiting for threads to catch up. If I close everything, i sometimes have to still wait from 30-45 seconds just to open the My Computer Icon. To me thats not acceptable. 26 processes running, with none larger than 10K, memory commit will go up to 90M of 1024M.. let its too slow to run anything.
Their support tell me, that even though its dedicated resources, i'm on a box with 4 other hosts., so hence it being slow. Their suggestion is I upgrade to another box.. I find it hard to believe that i should have to upgrade so I can run a few non-memory intensive programs.
What can i do to prove to them that there is a problem? I'm locked into a 3 month commitment. What use is a VDS server that run 1 application? Is there specific counters i can enable that would get to the heart of the matter?
It is Win2003 running on it. 60 G HD , 1.5 Ghz CPU, 1G Ram. I read some of the other threads, but they seem to be more for someone that has access outside of their VDS.. (unless i missed one)
edit: I suggested to them maybe someone is hammering the HD, but they say there is no hard disk, that its only allocated space..(?)
I've got 25 domains on a Virtuozzo/Plesk8.6/CentOS5 VPS. Each domain has one up-to-date install of WordPress, most have very little traffic (average 200mb per month), maybe 2 domains get 5-7gb traffic per month.
I monitor port 80 connections and rarely see more than 10 at a time. That should in my opinion be no problem at all for a VPS with 768mb guaranteed ram and 2.4ghz cpu. I've got 30gb hard drive spare too.
But.... about 8 or 10 times a day it grinds to a complete halt: server load at 500-1000%, sites timing out, plesk takes 3mins to load, often I can't even connect with SSH, and the plesk web server, apache
80 seconds sounds like a huge amount of time for a MySQL insert to me! Does anyone know if this is likely to be the cause of my trouble? Some problem with Plesk and the database? Or could it be something else?
I have always found this odd and do not understand enough to figure it out.
Any knowledgeable people would be cool.
My datacenter has given me 2amp circuit for my computer. It seems to run fine on it.
Looking at the back of my power supply it says the following
520Watt max
Input rating 100-240v~,50-60Hz 7A-3A
Output rating +3.3V & +5V combined 150W max 12V combined current 39A max (and a lot of other stuff on amps and volts for output).
Now, Watts = Volts * Amps. So if I have 520watts and 12 volts would be 43 amps If it means 520 watts and 3.3volts it would be 157 amps However it seems to max at 39amps on the back for output.
I imagine input is all that matters here and at 520 watts and 100 to 240volts that would be 5.2 amps to 2.16 amps
And the thing says 7amps to 3 amps on the back for input.
So....I imagine that is the max it can pull before dying.
However, I have 4 drives, 4 ram sticks, and 2 quad core cpus all going and I would figure I am at least halfway or more of the power use.
Main VPS is 1152mb SLM ram running only a relatively small vBulletin community. My goal is to keep this as fast as possible, which is why I overbought on the VPS and recently moved my couple other low volume sites (and some higher volume image hosting) + some email acccounts on to a 512mb SLM VPS.
Both VPS's are Cpanel/WHM running centros 4 (might be 4.5, I'm confused on this) and the latest release version of whm.
What I am wondering is what stats/tracking I might be able to run to see what typical usage is on the two VPS's for a day or two and try and figure out if I jumped the gun in splitting them, and can safely remerge them without effecting the vBulletin performance.
Can someone help me out with how to setup and read some stats that would tell me if both VPS's are underutilized enough to merge them back together onto the 1152mb VPS?
I'm trying to set up a site on IIS 5. I think I've done the setup on IIS ok, but when I try the site, it's not getting served. I think there might be a firewall in the way, blocking port 80. Any idea how I would test to see if this is the case?
Assuming a linux server / apache / php / mysql setup
I'm wondering 1) how you can find out which php script is causing problems, i assume infinite looping, crashing a server and 2) are there any measures you can take (maybe in the php or apache configuration) that can prevent such a thing from happening in the first place, other than writing good code obviously?
Is there anyway to determine how many visitors subscribe to an RSS feed? (besides using something like technorati or feedburner) Does anyone know of a log analyzer will do this or another custom solution?
I'm using MRTG to display graphs of traffic to my server, and today noticed a period of several hours where there was a lot of *incoming* data over the network, which is unusual. There's a good chance it's some harmless script or service that I've forgotten about, but really I need more information on what services/ports are creating what traffic, which brings me to my question.
I'm wondering whether there's an easy way to check what services (or ports) are creating traffic, and what volumes of traffic. It would be nice if it was possible to graph or breakdown the traffic by service somehow, and determine which are the biggest users of bandwidth and when. This is a dedicated CentOS server that I can install basically anything on, but that doesn't use a control panel, so I'd like to see what great ideas you folks can come up with, or whether there's any nice software you can suggest. I'm only using the server for hobbyist stuff at the moment, so free software, or ideas using built in commands are preferable.
a log file analysis program that can tell me the exact time an ip address accesses my server and how many times that particular ip address hits the server.
Does anybody know of a package that supports this, or is it even possible?
We're running on Linux/Apache/MySQL/RoR and have a number of cron jobs that run throughout the day on our server. We've been noticing lately that at certain times of the day the site becomes really slow. When I'm online with my engineers I can mention this to them and they can check and see and say "Oh yeah, it's job XYZ that's spiking the server load."
That's great but much of the time when I notice the sluggishness my developers are offline (we're in different time zones). I'm wondering if there's a fairly easy way to track this when they're not online so we can say "Yup, last night at 10 PM your time when you noticed that it was job ABC." There has to be something that allows you to do this right?
Cpanel installed on Centos 5.2 on a dedicated server.
This morning, I received multiple emails about services being restarted. About 20 minutes apart, with nothing that I coudl see going on with the server, I had multiple services fail and restart.
How do I trouble shoot this to see if there is a problem?
At 6:21 The following services failed: httpd lfd httpd cpsrvd lfd and I received emailed messages similar to these (obviously the port and service name were didfferent in each email). cpsrvd failed @ Sat Nov 15 06:21:29 2008. A restart was attempted automagically. Failure Reason: Unable to connect to port 2086 Then, at 6:42 The following services failed: cpsrvd cpsrvd httpd lfd httpd lfd and I received emailed messages similar to these (obviously the port and service name were didfferent in each email). cpsrvd failed @ Sat Nov 15 06:21:29 2008. A restart was attempted automagically.
I was told that I have bursted the RAM on my VPS, over 70,000 time since the parent server was rebooted 100 days ago. How do I determine what site is causing this (including mine, which isn't DB driven (yet))?
I am looking for a way to put a rewrite condition into my configuration where I can poll an external web page and based on the http status code, allow the rewrite or not.
I am doing this in order to set a maintenance status on two different proxy servers, one is apache the other is a big-ip. I already have the maintenance pages and control setup on the big-ip, but I need a way to poll a url and pull a specific http status code served up by the big-ip.
i have installed ppa 11.5 update #3 and now updated to update #4.
While adding new service nodes, i get everytime the error "Can not determine remote host "hostname" platform". I tryed Apache web server standalone and with mysql, postfix...
Whenever I restart my apache on my VPS, it throws some wierd warnings:
Forcing reload of web server (apache2)...apache2: apr_sockaddr_info_get() failed for goebs apache2: Could not reliably determine the server's fully qualified domain name, using 127.0.0.1 for ServerName waiting apache2: apr_sockaddr_info_get() failed for goebs apache2: Could not reliably determine the server's fully qualified domain name, using 127.0.0.1 for ServerName .
where goebs is my VPS linux machine hostname.
I already have a domain setup with this VPS, and it works fine, just that I get this weird warning.
At 15.00 PM (+1 hours here) my isp rebooted the hardware node. After the reboot there were some issues, they think it is possibly a DDoS Attack. They said the load on the host node was in the hunderds and my VPS was stopped because of the reboot and when I started it, HyperVM first told me it was not possible and after a few tries it started finally.
But the load was very high, and it still is. How can I check what the problem is that creates the high load?
I solve the sshd problem cos I edited the conf file with UseDNS no so it faster right now but I couldnt do anything with mc.
I have to wait about 15 seconds to load it.
Any suggestion?
or any ideal why my server become slower?
System: Debian 4.0 kernel: grsec 2.6.19-1
Till today morning its okey.
Oh, yes, I run tiger and got the following message in log:
09:50> Beginning security report for atdn.us (GNU/Linux Linux 2.6.19.1-grsec). # Performing check of passwd files... # Checking entries from /etc/passwd. --WARN-- [pass014w] Login (firebird) is disabled, but has a valid shell. --WARN-- [pass016w] User oident has / as home directory --WARN-- [pass015w] Login ID sshd does not have a valid shell (/usr/sbin/nologin). --WARN-- [pass015w] Login ID sync does not have a valid shell (/bin/sync). # Performing check of PATH components... --WARN-- [path009w] /etc/csh.login does not setenv an initial setting for PATH. # Only checking user 'root'
I modificated this to the correct one (I mean I try to solve this warnings) but when I noticed that my machine becmome slower, I recover everything.
I have an ASP.Net + SQL Server 2008 Express website I was hosting with DiscountASP.net. Their service seems pretty solid, and I didn't have any issues hosting with them. Customer support was quick and knowledgeable.
However, I had noticed for quite some time that the website seemed pretty slow. Please note that the website isn't live yet, so it was only me and a colleague who were accessing it. I was accessing the website from Dubai, UAE and he from Doha, Qatar. Therefore traffic/load definitely wasn't the problem.
We tried an instance of the website on a VPS (the DB was still running on DASP), and it was blazing fast! The difference in speed was quite significant and sustained. What could the reason for this be? Is it because we can't expect good speeds from shared hosting? Or should I report this issue to DASP?
I would be paying around 5x of what I'm paying right now ($100 instead of $20) if I switched to a VPS, but if there are definite advantages, I would be willing to do so. I don't expect large volumes of traffic on the website for quite a while.
