Is there anything I can do now?

rds100

MikHo said: Bumping tickets might put you at the bottom of the queue and delay responses.

While technically true, i doubt most of the unmanaged / self managed providers here have "queues" with tens of tickets waiting for reply. If they do have such queues, they must be doing something wrong.

MassNodes

@Profforg said:
Load 1 if you have 1 core means 100% load in most cases Load 38 if you have 1 core means something is definitely wrong (maybe you have been flooded).

I think you should try to contact your hosting provider with all available ways and tell to activate your VPS. However you must said, that such load will never happen again
After activation, connect with SSH and monitor your load, if it will be > 10 - you should dump some output of current state (ps aux, netstat, htop, e.t.c.) to text files and stop your web-server, database server and any other services which can cause such load. Then look at our txt dumps and look at that cause your load. If it's websites - you should see which user/website cause much load at these dumps. If it's database - you should look at binary logs and active userstat. I can look at that's the problem with your VPS, if you want. I am SysAdmin.

I get the feeling from your post that you assume load means CPU usage. It has to do with the amount of processes waiting to use the CPU on average for a specific time interval such as 1 min, 5 min and 15 min. If he's running a webserver that has cpanel with a lot of traffic (maybe what you consider a flood of requests) a lot of processes can get spawned and subsequently cause the load to become that high. Each httpd process has a limit of connections it will serve using it's own threads before a forked version of it will start to handle other connections.

mikho

@rds100 said:
While technically true, i doubt most of the unmanaged / self managed providers here have "queues" with tens of tickets waiting for reply. If they do have such queues, they must be doing something wrong.

I once had a vps with a provider that actually on purpose put your ticket "on hold" and added waiting time every time you bumped a ticket asking for a reply. Not the best option and that's one of the reasons I'm now a former customer and not a current customer.

So even if there is technically no queue the ticket wasn't answered.

sman

If you were causing a 38 load on my servers I would suspend you and probably ignore your tickets hoping you go away. Saying the limit is 1 with todays servers isn't unreasonable either. Anyone consistently going over about 1.5 usually goes on my radar.

sleddog

MassNodes said: I get the feeling from your post that you assume load means CPU usage. It has to do with the amount of processes waiting to use the CPU on average for a specific time interval such as 1 min, 5 min and 15 min.

Yes. And on Linux, load includes IO wait time. So dog-slow disk access can drive up your load by creating a queue of processes waiting for disk.

UltranetUSA

@sleddog said:

Tell that to google or Facebook or a telecom company, VPS is for kids or reaaaally low budget "business"

Profforg

MassNodes said: I get the feeling from your post that you assume load means CPU usage. It has to do with the amount of processes waiting to use the CPU on average for a specific time interval such as 1 min, 5 min and 15 min. If he's running a webserver that has cpanel with a lot of traffic (maybe what you consider a flood of requests) a lot of processes can get spawned and subsequently cause the load to become that high. Each httpd process has a limit of connections it will serve using it's own threads before a forked version of it will start to handle other connections.

Profforg said: in most cases

MassNodes

@sleddog said:
Yes. And on Linux, load includes IO wait time. So dog-slow disk access can drive up your load by creating a queue of processes waiting for disk.

Yup, your absolutely right. It includes uninterruptible sleep state processes.

sleddog

UltranetUSA said: VPS is for kids or reaaaally low budget "business"

Why do you come to a forum that is about using VPSs, and then belittle their usage? It makes no sense.

UltranetUSA

@sleddog said:

I sell VPS and Dedis and if you ask me I will recommend what is best for your business, I will never sell a VPS for a VOIP business etc, that will help avoid the case of our friend here in this tread!!

sleddog

I didn't say that....

Janevski

@serverian said:
Having load of >1 24/7 is considered high if you do not have dedicated cores.

^this

stallion

Alright!!!!! its all good now. My provider unsuspended my account. IN WHM I paid a visit to "resource monitor" and I was able to identify one account which was causing the huge load. Suspending that one cPanel account has fixed this problem. Now the load is around 2-4. Still high according to my provider's standards but I am happy it is not 38 anymore.
I was able to transfer all accounts out from there to my dedicated server.

It seems my VPS provider has hired some guy for support who answer's support tickets only once a day. But the replies I got from him were polite and as he restored my access to the VPS I am all happy with them. In fact I am thinking about writing a positive review for them on WHT although I nolonger need their service. It was my fault the load was jumping so high. I should have checked it earlier.

mikho

Glad it is all sorted and that you got all your files. Remember to keep your own backups in the future.

seaeagle

if you are using whm/cpanel i think the whmeasybackup is a great backup tool. get a leb here with penty of disk and run backups as often as you need them via cron.

ecircuit

my advise for the future is to use cloudlinux. from my experience it will save you a lot of headaches with abusive costumers.