Very High Load Averages

[SDH1]

After a very scrappy thread with Andre on the webhostingtalk forums, I've decided to try and get some answers here.

We've only been with DEHE for just over a month and our VPS - 263 - has had high load averages and a slow site for the past two weeks. Tech support are well aware of the issue - Ticket ID: NHV-94637 -and have tried several things to sort it out.

Yesterday, Andre told me that it was being caused by something that we've done in configuring the VPS, but I can't honestly remember having ever configured the VPS.

I emailed support three times yesterday to ask what we've done, but we've had no answer in about 18 hours.

This morning the load averages have got as high as 36, 21, 13

Our site is running at a crawl and it appears as though it has been doing so for about the last 9/10 hours.

Support have told me that they're working on it, but given no more info.

So please can you tell me what we've configured/misconfigured, why our site is running with such heavy load averages and when we can expect things to get back to normal?

[SDH1]

I've recently heard that today's very high load averages problem is due to Raid Initialisation and it's currently running at 61% complete - could you ellaborate at all?

Why is it initialising?
Will it cure the long term problems with the load averages?
Where was the post on the forums or support email to notify us that the system was having problems 11 hours ago?

Thanks

Stuart

[Milovan]

Stuart,

There was an emergency maintenance in the DC on couple of servers, among which one VPS node was. We had no time to email everyone what is going on, though we planned, however problems started piling up one after another. I was in charge over the node on which your VPS is located and everything was running smooth until somewhere around 4:50 AM, when RAID initialization started to malfunction and at that point, server had to be rebooted to have the init started again.

There was a small power outage last night in one room in the DC which lasted for no more than 5 seconds and which is the ultimate reason for the problems with this particular node. All clients who had opened tickets about this were exchanging emails with us in a matter of a minute or two, and at that very moment the thing that was most important is having the server stabilized.

I do apologize that you were not properly informed and I take all responsibility for that.

[SDH1]

OK, thanks for the info.

I take it that the power outage you mentioned is different to the one that tech support told me about the night before last, which gave our site a blackout for over an hour?

Could you also shed some light my other queries:

1. Any idea if this will now cure the long term problems with the load averages?

2. Yesterday, Andre told us that the load averaging problem was being caused by something that we've done in configuring the VPS, but I can't honestly remember having ever configured the VPS. Could you let me know what we've done?

3. Could you not post on the forums to notify us that the system was having problems 11 hours ago?

Thanks

Stuart

[ThunderStorms]

Same questions i ask here to you Stuart:

Have you been in virtuozzo?

If so have you checked all logs etc.?

If not maybe there you can find a problem/change/solution etc.

[SDH1]

Hi wavelandsonline, I take it that you're WOS on the webhostingtalk forums?

I've been in the Virtuozzo CP and looked at the 'Action Logs', but there is no log before 6th January and we first reported the problem on the 4th.

Should I be checking anywhere else in the CP?

Thanks

Stuart

[ThunderStorms]

Hello Stuart,

Yeah that's me on WHT.

Have you checked the QoS alerts also ?

And the traffic log?

Btw sorry i couldn't reply earlier as both of my personal systems crashed at same time, lol.

Regards,

Arno

[SDH1]

The QoS is only showing one log from a couple of days ago and I think it coincides with the datacentre's power outage.

The traffic logs look Ok and the status changes are only showing starting, stopping, running and down for the past 24 hours or so.

Still can't find anything relating to us making changes before 4th January

[lunoto]

Stuart,

There was an emergency maintenance in the DC on couple of servers, among which one VPS node was. We had no time to email everyone what is going on, though we planned, however problems started piling up one after another.

Does the nice Dehe support team is still busy ?
I need to have some advices from them.
Or is it better to wait a bit ?

[Andre]

No all problems have been resolved. Please feel free to submit a support ticket at any time.

[Milovan]

Does the nice Dehe support team is still busy ?
I need to have some advices from them.
Or is it better to wait a bit ?

Just want to follow up on this. Even though there were emergencies, I don't feel that any of the customers who's VPS' were not affected by this experienced lack of assistance from our end during this time.

[SDH1]

Andre, could let you let me have some info on the questions I posted in #4

Our VPS is running well at the moment with low averages , but

Yesterday, you told us that the load averaging problem was being caused by something that we've done in configuring the VPS, but I can't honestly remember having ever configured the VPS. Could you let me know what we've done so we don't do it again.

Could you also post faults as they occur on the forums to notify us that the system is having problems?

Thanks

Stuart

[Milovan]

Stuart,

Andre had to leave the office, but he will be back in a couple of hours. I'll leave him a note to get back to you.

[Andre]

Hi Stuart,

I didn't mean to say that you misconfigured anything yourself, but for a website with high traffic it's always highly recommended to apply a custom configuration for Apache, PHP and MySQL. For instance with eAccelerator we have achieved incredible results in the past (for instance a customer's server which always had a load average of about 30, and after we had tweaked the configuration the load was always under 1).

What I meant to say was that the high load in your VPS (before the power outage) was not caused by our servers, but rather by the configuration of those services. I believe our techs have already installed eAccelerator for you, which should help a lot. In case the load still gets high we can try to analyze your sites and see whether the configurations can be tweaked any further.

I hope that answers your question. If you have any additional questions, feel free to ask any time.

[SDH1]

Hi Stuart,

I didn't mean to say that you misconfigured anything yourself

But you virtually underlined that fact on the webhostingtalk forum. Anyone reading that would yet again make the assumption that you were accusing us for causing the problem

The only thing that's happening is that his site was slower than he would expect it to be. That has to do with the way his VPS is configured (that is: the way he configured it before he requested troubleshooting), and we helped him about as much as we possibly could have to get things running as smoothly as possible.

I hope that you will now be making an apology on the WHT forums, to rectify your comments.