Is this server load due to node overselling / abuse?

kauriman

Hi everyone!
I've been with chicagoVPS for over an year now, got a deal for an OVZ
3GB ram
100GB HDD
I'm hosting 4 Wordpress sites with help of EasyEngine, Nginx, w3totalcache. Not too much traffic, all of them with CloudFlare and Photon CDN
Lately I've been experiencing some downtime and high server load.

I'm concerned because there is little CPU and MEM usage and still high load on top

Is this normal?

:~# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 39.8786 s, 26.9 MB/s

26.9 MB/s I/O is allright?

Do you think this is due to node abuse or overselling?

I know, I'm a noob, but doing my best to learn.
Any help is much appreciated!

maldovia

That IO is very slow. But honestly for $3/mo or whatever you're paying.. what do you expect?

root@test:~# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 1.75318 s, 612 MB/s

That is what I get with my test VPS at VPSHaus, should give an idea of what IO should look like.

It's pretty well documented at this point that CVPS probably oversells, but again for $5 for an 8GB VPS what can you really expect.

kauriman

Thanks for your words maldovia, do you think the 2.03 server load is due to IO? It even got to 10 and over.

maldovia

@kauriman said:
Thanks for your words maldovia, do you think the 2.03 server load is due to IO? It even got to 10 and over.

No - I think it's because its oversold.

kauriman

Rigth! so OVZ top is telling me the node's load

Will KVM virtualization isolate me and show me actual container's load?

hostdare

Maybe open a ticket with them also ?

kauriman

Yeah I did so here's their reply:

Hello x,

I am still monitoring your server and working fine at my end if I found downtime will let you know about server performance.

Thank you,

Sarah S.
Technical Support,
ChicagoVPS.

on a previous message they told me there was something going wrong

We have taken the strict action against on the thous server which was increasing the load on server that may caused you did faced issue,

Now your server is working fine, don't worry you will never face downtime issue again:

no need to move your server to another nod let us know if you need further assistance..

Sarah S.
Technical Support,
ChicagoVPS.

But the truth is that is happening over and over again

MasonR

Now your server is working fine, don't worry you will never face downtime issue again.

That's a bold statement.

maldovia

@MasonR said:

Now your server is working fine, don't worry you will never face downtime issue again.

That's a bold statement.

"Sarah S."

"Dedicated to Uptime"

kauriman

I know, I must move out from them ASAP, its a shame I deposited money in advance because their paypal subscription wasn't working at all.

Been trying to get a KVM slice from BuyVM but their fraud system wont let me, I've even talked with Fran with no luck!

I'm kinda stuck here

AuroraZ

Try VirchMach they are a hell of a lot better then where you are currently. They seem to be good guys and have a good service for the price.

K4Y5

@maldovia said:
That IO is very slow. But honestly for $3/mo or whatever you're paying.. what do you expect?

root@test:~# dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 1.75318 s, 612 MB/s

That is what I get with my test VPS at VPSHaus, should give an idea of what IO should look like.

It's pretty well documented at this point that CVPS probably oversells, but again for $5 for an 8GB VPS what can you really expect.

You forgot to mention that part that the IO is from a test VPS on a test node with NVMe drives instead of conventional HDDs or SSDs.

Not that I have anything nice to say about ChicagoVPS since they oversell nodes as a part of their business policy and I just lost a VPS because their invoices never made it to my inbox. But, I still feel that apples should be compared with apples and not oranges.

@kauriman that IO performance is not too shabby considering its a HDD based VPS based on OVZ - which is usually oversold like there is no tomorrow.

@kauriman said:
Yeah I did so here's their reply:

Hello x,

I am still monitoring your server and working fine at my end if I found downtime will let you know about server performance.

Thank you,

Sarah S.
Technical Support,
ChicagoVPS.

on a previous message they told me there was something going wrong

We have taken the strict action against on the thous server which was increasing the load on server that may caused you did faced issue,

Now your server is working fine, don't worry you will never face downtime issue again:

no need to move your server to another nod let us know if you need further assistance..

Sarah S.
Technical Support,
ChicagoVPS.

But the truth is that is happening over and over again

I doubt that's a 'Sarah' based in the US of A, considering the gibberish that 'she' responded to your ticket with. CVPS should at-least outsource support to people who can write cohesive / coherent fucking sentences while responding to tickets.

Francisco

kauriman said: Been trying to get a KVM slice from BuyVM but their fraud system wont let me, I've even talked with Fran with no luck!

Try a PM

Let me know what's up, or send an email to [email protected]

Francisco

FredQc

K4Y5 said: You forgot to mention that part that the IO is from a test VPS on a test node with NVMe drives instead of conventional HDDs or SSDs.

[root@vir ~]#  dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 1.1168 s, 961 MB/s

[root@vir ~]# uptime
 22:46:35 up 276 days, 11:55,  1 user,  load average: 0.08, 0.07, 0.06

Virmach $5/year

:P

teamacc

@kauriman said:
Yeah I did so here's their reply:

which was increasing the load on server that may caused you did faced issue,

@wss i found something for you to put in your signature.

AnthonySmith

Well 3 things.

1. What real world impact is it having?

2. Ignore 99% of these IO tests, they are sequential write only and 99/100 people are hitting a raid card cache or its running in ram through ploop, this is the equivalent of people thinking that if the car speedo shows 100 mph that they are doing 100mph even if the car is not even moving.

3. The load really is not that high, I suspect the real cause here is that you are heavy on mysql random reads/writes and you have noisy neighbors also generating a lot of random IOPS, the clue being the .wa

But really, if your sites load OK 95% of the time, done worry about it, your paying less than a bag of Doritos costs for an entire month of service, if they are not loading fine and you are prepared to pay more, go for it.

WSS

@teamacc said:

@kauriman said:
Yeah I did so here's their reply:

which was increasing the load on server that may caused you did faced issue,

@wss i found something for you to put in your signature.

I saw this yesterday and had a pretty good snicker to myself. Even in jest, nothing CVPS or HVH goes in my sig.

kauriman

@Francisco said:

kauriman said: Been trying to get a KVM slice from BuyVM but their fraud system wont let me, I've even talked with Fran with no luck!

Try a PM

Let me know what's up, or send an email to [email protected]

Francisco

Sent you PM & email

Francisco

@kauriman said:

@Francisco said:

kauriman said: Been trying to get a KVM slice from BuyVM but their fraud system wont let me, I've even talked with Fran with no luck!

Try a PM

Let me know what's up, or send an email to [email protected]

Francisco

Sent you PM & email

And pinged back. Karen will take care of ya'.

Thanks for the interest and future signup!

Francisco

sin

There's so many more providers that you can be with for around the same price and will offer you better performance/reliability.