VPS and resources - UK Web Hosting | Dedicated Server Windows and Linux VPS Forum

**martynh99** · #1 (**permalink**) 24-01-2008, 16:57

Ok, I'm on a VPS Hosting running on the new servers and it's been running like a 3 legged dog.

Support now tell me that 'There was a VPS Hosting on main server which were creating problems on the main server. We have stopped and suspended them and now everything seems to be running fine now.'

To quote from the sales pages on Windows Reseller 'We do NOT allow overselling on any of our reseller hosting plans, however you can consider our VPS Hosting or Semi-Dedicated hosting plans as they offer flexibility to oversell and keep CPU & RAM usage within acceptable performance metrics.'

So what went wrong ?

The performance of my VPS Hosting was not acceptable and I only had 2 sites running at the time, 1 small Vbulletin instance and another MySql based web site hosting that was not been used at the time as i've only just transferred the software across, it's not even configured yet.

How can another VPS Hosting have such a detrimental effect on my VPS Hosting ?
This isn't what I signed up for

**gobold** · #2 (**permalink**) 24-01-2008, 22:49

I must be on the same Node as you cos 1 of my Vps was off for 23 hours and I got the same message. My other VPs was on the same node and also crashed.

As I write they are now playing up. Live chat insist they are fine but me and my daughter (8 miles away) cant access.

Ummmmm

**martynh99** · #3 (**permalink**) 24-01-2008, 23:23

I just got this reply from tech support about how VPS Hosting works

Quote:

'Yes, if you will exceed the allocated resources then we gets the notification then we suspends your VSP and notifies you. We are having our own monitoring tools so other VPS Hosting will definitely not affect to you.'

Not quite sure what the section in blue means. Do you have the monitoring tools in place or are you getting some monitoring tools some time in the future ?

If they are already in place, why have I encountered 2 serious elongated performance problems that have only been resolved when I have gone first through live chat to be told that there is no problem, then onto support who say

Quote:

We have investigated and made some changes on the sever

or

Quote:

There was a VPS Hosting on main server which were creating problems on the main server. We have stopped and suspended them and now everything seems to be running fine now.

This VPS Hosting stuff sounds really dodgy to me. You sign upto a package that is limited to 7gb disk , 512mb memory.

It then transpires it's not limited but it is upto you to stay within these limits or your VPS Hosting will get suspended (see section in red above)without warning.

Again I quote the sales blurb on your website hosting

Quote:

consider our VPS Hosting or Semi-Dedicated hosting plans as they offer flexibility to oversell and keep CPU & RAM usage within acceptable performance metrics

- so you have the flexibility to oversell but it's upto me to monitor my CPU & RAM 24*7 or you will probably suspend me

**gobold** · #4 (**permalink**) 24-01-2008, 23:33

Martyn

You may want to look at my thread bout my problems on the node at the weekend
VPS DOWN - In desperation I write here

I have other hosting and no problems like this.

I would read my thread cos there has been a good response from eukhost and I am hopefully they will get to the bottom of it and sort it

**martynh99** · #5 (**permalink**) 24-01-2008, 23:46

Another quote to get your teeth round, this time from

http://www.eukhost.com/<a href="http...guidelines.php

Quote:

II. A Windows VPS Hosting get "Guaranteed RAM" and "Equal Share Of CPU".
-> Equal CPU Share - The VPS Hosting gets to share any available CPU over and above their allocated CPU resources, if available. However, the share depends on the plan you are on. A W-VPS Hosting-03 will always get a larger share than a W-VPS Hosting-01.

So I have an equal share of the CPU's on the 'big' server and 512mb guaranteed ram - tell me again how some other VPS Hosting impacted me when I was nowhere near my limits. When I suffered the problems, the VPS Hosting was running around 280mb max of RAM used and CPU's around 30% max peak.

**martynh99** · #6 (**permalink**) 24-01-2008, 23:48

And just one more

http://www.swsoft.com/r/pdfs/vz/whit...Z-Overview.pdf

If you read the section titled 'dynamic resource allocation' on page 6 you will see that you can set limits per VE

Again if this is in place, no other VPS Hosting could affect mine.

**gobold** · #7 (**permalink**) 24-01-2008, 23:58

Having been in the biz a while and having been on a server that was hit by a DOS attack it can effect us.

If VPS Hosting/Server etc is running bad scriipts etc or trying to do bad things, usually hacker etc, this can have a effect on the overall node. I have seen this.

I server I was on keep getting speed problems and it was just run away scripts, another (servage) had a account set up for attack purposes and it caused major problems.

**martynh99** · #8 (**permalink**) 25-01-2008, 00:11

Quote:

Originally Posted by gobold

If VPS Hosting/Server etc is running bad scriipts etc or trying to do bad things, usually hacker etc, this can have a effect on the overall node. I have seen this.

But my point is if the node is setup correctly to limit the maximum resources a VPS Hosting can take, then a VPS Hosting being attacked cannot affect the main node.

The quote the support team made shows that the limits are not in place as if a VPS Hosting exceeds it's plan it's shutdown without warning

**eUKhost.com** · #9 (**permalink**) 25-01-2008, 00:23

We have restricted our Windows VPS Hosting plans to specific resource limits, like 100 process limit for W-VPS Hosting-01, 150 for next and so on. These limits don't make much difference as even 1 process is sufficient to overload a server.

One option ahead of us is to enable auto-shutdown in Virtuozzo which will automatically shutdown any VPS Hosting that tries to overload, but this feature has created more problems in past as VPS Hosting's were shutdown frequently and customers cannot reboot their VPS Hosting if Virtuozzo suspends their VPS Hosting.

There's no other way to rectify this problem so we have started adding new nodes with Dual Quad Core Xeons to improve processing power of the nodes.

**martynh99** · #10 (**permalink**) 25-01-2008, 01:13

Quote:

Originally Posted by eUKhost.com

We have restricted our Windows VPS Hosting plans to specific resource limits, like 100 process limit for W-VPS Hosting-01, 150 for next and so on. These limits don't make much difference as even 1 process is sufficient to overload a server.

Fully appreciate this but the sales blurb I quoted earlier states you have hard limits set on Windows VPS Hosting - if this was the case, the 1 rogue process would hit this limit and be limited thus not affecting the other VPS Hostings.

Quote:

One option ahead of us is to enable auto-shutdown in Virtuozzo which will automatically shutdown any VPS Hosting that tries to overload, but this feature has created more problems in past as VPS Hosting's were shutdown frequently and customers cannot reboot their VPS Hosting if Virtuozzo suspends their VPS Hosting.

Again if each VPS Hosting has a hard limit, problem goes away.

Quote:

There's no other way to rectify this problem so we have started adding new nodes with Dual Quad Core Xeons to improve processing power of the nodes.

Yep and I'm on one of these new nodes and i'm having repeat problems.
It's also quite clear that any monitoring you have in place is not sufficient as on both occasions, 1st line support have said there was no problem on my VPS Hosting. Surely they would check the physical node aswell for any issues - wouldn't they ?

As i've mentioned many times before, I moved to EUKHost as I was unhappy with the slapdash approach to support that my previous host had - a MySql server that kept overloading connections and they could never see the problem. To fix it they would just restart it, not address the underlying problem - i'm seeing a trend in the approach to my problems. First tell me there isn't a problem, then tell me it's fixed when it isn't - the email prob on Weds eve being a prime example of this. I basically spelt out what the problem was but because a test email went through to a gmail.com account it was deemed fixed.

I have over 20 years IT experience, 15 years of that implementing & supporting solutions on the NT platform and its successors for large blue chips and got my 1st MCSE 12 years ago - I know my systems so I can't be fobbed off.

**gobold** · #11 (**permalink**) 25-01-2008, 10:26

I can see where you are coming from. I agree with what you say.

I am at the limit of my patience at eukhost. But if was was somewhere else my patience would have expired by now. I am gonna wait a lilttle longer to see if this problem is fixed as I think they may well sort it.

Yes they are slow sometimes and things are not clear. BUT I have see a real passion and desire from the top to fix things and try to make the company the best they can. Thats why I shall wait a little longer.

**jc8654** · #12 (**permalink**) 25-01-2008, 10:55

I'll point you to this thread which is about what's causing a load of problems. Obviously with having to do this changeover, it's putting more pressure onto the support department as I'm sure you can appreciated. Extra people have been moved into the VPS Hosting department to help solve the issue and eUK are recruiting for extra staff to resolve the issue permanently.

Anyway, the link: VPS Issues - Migration

**martynh99** · #13 (**permalink**) 25-01-2008, 11:14

Quote:

Originally Posted by jc8654

I'll point you to this thread which is about what's causing a load of problems. Obviously with having to do this changeover, it's putting more pressure onto the support department as I'm sure you can appreciated. Extra people have been moved into the VPS Hosting department to help solve the issue and eUK are recruiting for extra staff to resolve the issue permanently.

Understand totally about your resourcing problems during the migrations but most of the issues i've been having have been down to simple config errors - things I would have expected to be templated in an environment such as yours.
Fundamental things such as changing IP addresses bound to a VPS Hosting when you move it to a new node, even when this was pointed out the MySql was left bound to the old IP's - in my view silly mistakes that could easily be avoided with a level of QA or simple checklisting being in place.

I still have my old hosting in place and that's where the bulk of my domains are still running, i'm going to hold out a few more days and see how the VPS Hosting pans out before making the decision on whether to move the remaining ones or ditch the VPS Hosting here.

**eUK-Nick** · #14 (**permalink**) 25-01-2008, 11:25

Quote:

Originally Posted by martynh99

Understand totally about your resourcing problems during the migrations but most of the issues i've been having have been down to simple config errors - things I would have expected to be templated in an environment such as yours.
Fundamental things such as changing IP addresses bound to a VPS Hosting when you move it to a new node, even when this was pointed out the MySql was left bound to the old IP's - in my view silly mistakes that could easily be avoided with a level of QA or simple checklisting being in place.

I still have my old hosting in place and that's where the bulk of my domains are still running, i'm going to hold out a few more days and see how the VPS Hosting pans out before making the decision on whether to move the remaining ones or ditch the VPS Hosting here.

Hello Martyn,

I can understand your concerns. I am having your VPS Hosting checked even as I write this post. Your ticket would be updated with the progress.

**eUK-Nick** · #15 (**permalink**) 25-01-2008, 15:28

Hello Martyn,

Your ticket has been replied to now.

Just for the benefit of the forum members i would like to share some information.

Presently, we are using the UBC (user bean counters). There are a wide range of options & parameters to be considered in this system. UBC also does not have any parameter corresponding to an intuitively clear concept of total memory size. As a result, UBC impedes Virtuozzo adoption. UBC presents the notion of a Guaranteed and Burst RAM, but at the same time there are 20 different memory related variables such as maximum processes, open files, iptables entries, kernel objects, disk cache, and TCP sockets that are limited independently of the RAM values and these vary greatly. If the UBC limits are set too low you will never actually be able to utilize your full RAM allocation.

We are once again going to test the SLM mechanism on our new nodes. If we get the results & performance which we want to derive, then we dont see any reasons of not having it implemented on the live nodes. We would however require more time to test SLM fully on the new nodes.

If & when we do implement SLM, you would surely be given the option of moving to that node. However, I do not have a time-frame for this.

**martynh99** · #16 (**permalink**) 26-01-2008, 22:37

Ok, as I have been posting in here as I had problems, it's only fair that I post when everything is ok

or it would look extremely biased.

I can say that the last 24 hours has been problem free, i've now got 1 live domain and 2 test domains on there. I'll be looking to migrate some more this week and also setup another new domain on there.

Thanks

for your efforts over the last few days, hopefully we're over the initial hurdles now.

**gobold** · #17 (**permalink**) 26-01-2008, 22:59

Afreed, everythig on my 2 VPS Hosting's has also been fine.

As with you lets hope thats sorted.