KVM VPS | 8 Locations | From £16 Per Year | DDOS Protected.

dahartigan

@dahartigan said:

@uptime said:
Can it get any crazier???

I think the Doc's been getting into the medicine cabinet lol

@HostDoc said:

@FAT32 said:
@HostDoc Just curious, is there by any chance at least one of the founders is actually a doctor?

Yes, my imaginary friend..

Confirmed

uptime

If your browser has been serving that page all this time, you have missed a lot.

LOL, that's probably just as well.

(And good to know that 6 GB deal was indeed a rare breed. A beast of an offer. to be sure.)

cybertech

6GB a production beast. Backup sleeper beast in Germany 4x4.

https://browser.geekbench.com/v4/cpu/12152030

dahartigan

The 6GB beast in production:

----------------------------------------------------------------------
CPU model            : Intel(R) Xeon(R) Gold 6128 CPU @ 3.40GHz
Number of cores      : 3
CPU frequency        : 3392.026 MHz
Total size of Disk   : 98.0 GB (12.8 GB Used)
Total amount of Mem  : 5960 MB (4041 MB Used)
Total amount of Swap : 1022 MB (554 MB Used)
System uptime        : 6 days, 7 hour 7 min
Load average         : 0.19, 0.20, 0.27
OS                   : Debian GNU/Linux 9
Arch                 : x86_64 (64 Bit)
Kernel               : 4.15.18-11-pve
----------------------------------------------------------------------
I/O speed(1st run)   : 786 MB/s
I/O speed(2nd run)   : 897 MB/s
I/O speed(3rd run)   : 912 MB/s
Average I/O speed    : 865.0 MB/s
----------------------------------------------------------------------

uptime

@dahartigan said:
The 6GB beast in production:

giggity

EDIT2:

That 1 GB dedicated swap ram partition space - NVMe I presume?

(Notice you manage to use 4 GB ram plus another 554 MB swapped out. Nice touch - I'm guessing this reflects your nested containers? I hope they are all idling, as is tradition, but whatever - I'll allow it.)

@HostDoc seriously, well done on that one, all around high quality for the Dallas node (and here's hoping you can keep it that way as the node gets more use.)

teamacc

@dahartigan said:
production:

Load average         : 0.19, 0.20, 0.27

dahartigan

@teamacc said:

@dahartigan said:
production:

Load average         : 0.19, 0.20, 0.27

root@tx:~# stress --cpu 3
stress: info: [7219] dispatching hogs: 3 cpu, 0 io, 0 vm, 0 hdd

And while that's going:

----------------------------------------------------------------------
CPU model            : Intel(R) Xeon(R) Gold 6128 CPU @ 3.40GHz
Number of cores      : 3
CPU frequency        : 3392.026 MHz
Total size of Disk   : 98.0 GB (12.8 GB Used)
Total amount of Mem  : 5960 MB (3936 MB Used)
Total amount of Swap : 1022 MB (726 MB Used)
System uptime        : 6 days, 8 hour 21 min
Load average         : 1.43, 0.80, 0.58
OS                   : Debian GNU/Linux 9
Arch                 : x86_64 (64 Bit)
Kernel               : 4.15.18-11-pve
----------------------------------------------------------------------
I/O speed(1st run)   : 838 MB/s
I/O speed(2nd run)   : 841 MB/s
I/O speed(3rd run)   : 996 MB/s
Average I/O speed    : 891.7 MB/s
----------------------------------------------------------------------

dahartigan

@uptime said:

@dahartigan said:
The 6GB beast in production:

giggity

EDIT2:

That 1 GB dedicated swap ram partition space - NVMe I presume?

(Notice you manage to use 4 GB ram plus another 554 MB swapped out. Nice touch - I'm guessing this reflects your nested containers? I hope they are all idling, as is tradition, but whatever - I'll allow it.)

@HostDoc seriously, well done on that one, all around high quality for the Dallas node (and here's hoping you can keep it that way as the node gets more use.)

I can't work out why it's decided to swap 554mb, but what I do know is Proxmox does some crazy shit from time to time.

uptime

proxmox does what it gotta do - and then some, I reckon

teamacc

@dahartigan said:

@uptime said:

@dahartigan said:
The 6GB beast in production:

giggity

EDIT2:

That 1 GB dedicated swap ram partition space - NVMe I presume?

(Notice you manage to use 4 GB ram plus another 554 MB swapped out. Nice touch - I'm guessing this reflects your nested containers? I hope they are all idling, as is tradition, but whatever - I'll allow it.)

@HostDoc seriously, well done on that one, all around high quality for the Dallas node (and here's hoping you can keep it that way as the node gets more use.)

I can't work out why it's decided to swap 554mb, but what I do know is Proxmox does some crazy shit from time to time.

Output of cat /proc/sys/vm/swappiness?

dahartigan

@teamacc said:

@dahartigan said:

@uptime said:

@dahartigan said:
The 6GB beast in production:

giggity

EDIT2:

That 1 GB dedicated swap ram partition space - NVMe I presume?

(Notice you manage to use 4 GB ram plus another 554 MB swapped out. Nice touch - I'm guessing this reflects your nested containers? I hope they are all idling, as is tradition, but whatever - I'll allow it.)

@HostDoc seriously, well done on that one, all around high quality for the Dallas node (and here's hoping you can keep it that way as the node gets more use.)

I can't work out why it's decided to swap 554mb, but what I do know is Proxmox does some crazy shit from time to time.

Output of cat /proc/sys/vm/swappiness?

60

teamacc

@dahartigan said:

@teamacc said:

@dahartigan said:

@uptime said:

@dahartigan said:
The 6GB beast in production:

giggity

EDIT2:

That 1 GB dedicated swap ram partition space - NVMe I presume?

(Notice you manage to use 4 GB ram plus another 554 MB swapped out. Nice touch - I'm guessing this reflects your nested containers? I hope they are all idling, as is tradition, but whatever - I'll allow it.)

@HostDoc seriously, well done on that one, all around high quality for the Dallas node (and here's hoping you can keep it that way as the node gets more use.)

I can't work out why it's decided to swap 554mb, but what I do know is Proxmox does some crazy shit from time to time.

Output of cat /proc/sys/vm/swappiness?

60

Means your vm starts swapping at 40% memory usage. You can lower it (and hence use less swap) by setting vm.swappiness=10 in your /etc/sysctl.conf

dahartigan

@teamacc said:

@dahartigan said:

@teamacc said:

@dahartigan said:

@uptime said:

@dahartigan said:
The 6GB beast in production:

giggity

EDIT2:

That 1 GB dedicated swap ram partition space - NVMe I presume?

(Notice you manage to use 4 GB ram plus another 554 MB swapped out. Nice touch - I'm guessing this reflects your nested containers? I hope they are all idling, as is tradition, but whatever - I'll allow it.)

@HostDoc seriously, well done on that one, all around high quality for the Dallas node (and here's hoping you can keep it that way as the node gets more use.)

I can't work out why it's decided to swap 554mb, but what I do know is Proxmox does some crazy shit from time to time.

Output of cat /proc/sys/vm/swappiness?

60

Means your vm starts swapping at 40% memory usage. You can lower it (and hence use less swap) by setting vm.swappiness=10 in your /etc/sysctl.conf

Thanks for the advice. Swapping isn't something that bothers me the way it seems to bother others.

From what I understand Linux has method to the madness when it comes to swapping, and 60 is a good compromise.

If I find that my loads are too high due to swapping, I'll download more RAMs or just ship the offending RAM consumer somewhere else

I appreciate your concern and your advice though

cybertech

I thought we were past swap

FAT32

@cybertech said:
I thought we were past swap

Swap is never enough

AlwaysSkint

With Centos 7/WHM/cPanel/csf:

nench.sh v2019.02.20 -- https://git.io/nench.sh

benchmark timestamp: 2019-02-24 11:42:34 UTC

Processor: Intel(R) Xeon(R) Gold 6128 CPU @ 3.40GHz
CPU cores: 3
Frequency: 3392.026 MHz
RAM: 3.7G
Swap: 2.0G
Kernel: Linux 3.10.0-957.5.1.el7.x86_64 x86_64

Disks:
vda 30G HDD
vdb 60G HDD

CPU: SHA256-hashing 500 MB
1.807 seconds
CPU: bzip2-compressing 500 MB
6.104 seconds
CPU: AES-encrypting 500 MB
1.445 seconds

ioping: seek rate
min/avg/max/mdev = 100.9 us / 179.7 us / 8.50 ms / 156.3 us
ioping: sequential read speed
generated 12.1 k requests in 5.00 s, 2.96 GiB, 2.42 k iops, 605.8 MiB/s

dd: sequential write speed
1st run: 454.90 MiB/s
2nd run: 661.85 MiB/s
3rd run: 709.53 MiB/s
average: 608.76 MiB/s

IPv4 speedtests
Cachefly CDN: 203.12 MiB/s
Leaseweb (NL): 18.50 MiB/s
Softlayer DAL (US): 91.31 MiB/s
Online.net (FR): 13.24 MiB/s
OVH BHS (CA): 23.91 MiB/s

AlwaysSkint

Reducing swappiness can lead to better performance - less I/O. I usually set 1 or 5, depending on my mood/weather/moon cycle/HostDoc crazy Flash deals. For SSD in particular, you do want it low, from a drive wear perspective, if nothing else.

Oh no! It's swap shop again! :-o

dahartigan

@AlwaysSkint said:
Reducing swappiness can lead to better performance - less I/O. I usually set 1 or 5, depending on my mood/weather/moon cycle/HostDoc crazy Flash deals. For SSD in particular, you do want it low, from a drive wear perspective, if nothing else.

Oh no! It's swap shop again! :-o

I will do that, you sold me on the performance and drive wear.

AlwaysSkint

@dahartigan
Just checked I have 1 and 5 set on each of my Proxmox nodes.

On the one set at 1:

free -m
total used free shared buff/cache available
Mem: 7876 3793 351 241 3730 3536
Swap: 4094 93 4001

The other (with excessive swap space)

Swap: 7167 538 6629

Goes off to tweak, if I can find time.

EDIT 2: Now sized to 2GB

imok

@HostDoc said:

@imok said:
When you say "USA node" you really say you have only one node there?

Just the one for now hoping to add LA soon but that has been delayed.

I wanted to buy the same 2 packages and have it in a different physical machine, but not a problem. Hope you can expand soon.

EDIT2:

@uptime said:

Seems like the doc just went crazy.

that 10 TB (traffic) 6 GB ram 3 cores 50 GB NVMe 50 GB HDD in Dallas though ...

So I disconnect a couple of hours and they decide to go crazy.

EDIT3:
OK, I'm a human that needs to sleep at night.

AlwaysSkint

@imok said:
OK, I'm a human that needs to sleep at night.

Over-rated - it's an age thang!

cybertech

Do notify if LA comes.

andrew1995

Where is the 6GB ram offer? Am I late? I am refreshing this page at least every 3 hours but no 6Gb ram.

dahartigan

Who needs sleep when there is the threat of missing out on a deal?

cybertech

It is during this flash sales that separates mortals and zombies

AlwaysSkint

Wish I held off now, seeing the ones with larger storage. Buyer's remorse? Nah.

uptime

@andrew1995 sorry the 6 GB is a mythical creature of yore

it has come, then gone, to be seen no more

Now is only whispered about in feverish ramblings, on random threads

An old whaler, peg legged and reeling, ranting:

Bitch get out of my head!

eol

I recommend swapoff --all.

iHavenoName

@HostDoc said:
USA node in located in Dallas and our DC provider is Incero.

FYI, this is the same DC as ssdnodes and I can confirm my hosting with them is a great location and blazing speed. This deal intrigued me, because it's about 50 usd cheaper (I don't need all the ram ssdnodes offers, but can use more drive space).

Oh and golf clap for the epic fucking domain hack: https://hostdochost.i.ng/pricing

Neoon

Sadly, the FIN box seems to have a overloaded uplink, which makes it not really usable with smokeping. Even over hours.

I get to many false reports the recent days, so another 25EUR burned, sad.