Fire At OVH Canada

FourOh-LLC

Placekey: 22k@5sb-8hm-54v - - - PlusCode: 86JH3PV5+PQ, Wed, 10 Mar 2021 04:51:42 -0800 last edited: Wed, 10 Mar 2021 04:52:34 -0800

This is something I just read on Gab - but what comes to mind first is they set the fire themselves. Whatever is the cause, maybe to blame some political opponent for it? Anyhow, I live a thousand miles from Europe, and 99% of the "new" is going to be clueless, deception, sensational, etc. etc. more of the same, more of the same.

Our new Secondary, zot.factoryfouroh.net is hosted on OVH Canada, which is not directly affected.
I am still looking for a LOCAL host outside of US and Canada, that is not operated by one of these giants, I am not interested in 1&1 IONOS, GoDaddy Europe, Linode Germany, etc.. etc..

I would like to spread the FourOh-LLC name servers and directory servers to 3 or more totally independent networks, preferably on different continents.

Einer von Vielen

Wed, 10 Mar 2021 05:01:38 -0800

Have you looked at this provider https://www.netcup.de/

Max Kostikov

Wed, 10 Mar 2021 05:28:25 -0800

@Einer von Vielen (einer@misty.digitalesparadies.de)

Have you looked at this provider https://www.netcup.de/

If you use it can you test disk performance? I'm interested how many iops they allow on VPS.

FourOh-LLC

Wed, 10 Mar 2021 05:10:47 -0800

I have a bias against anything hosted or coming out of Germany, and from Western Europe in general. I prefer something owed inside Central or Easter Europe,Italy, even Greece. India, South Korea, Japan, Australia. Some place which has at least some fight left in them against US and EU laws.

FourOh-LLC

Wed, 10 Mar 2021 05:53:13 -0800

Here is the "official" development tracking over the fire, sent to my account by OVH.
OVH Tasks

OVH RSS Feed

Einer von Vielen

Wed, 10 Mar 2021 06:45:12 -0800

If you use it can you test disk performance? I'm interested how many iops they allow on VPS.

How?
(I have root on the terminal.)

Max Kostikov

Wed, 10 Mar 2021 07:42:32 -0800

@Einer von Vielen (einer@misty.digitalesparadies.de)

How?

You may install tool fio and then run test with something like


fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randread

Einer von Vielen

Wed, 10 Mar 2021 06:47:27 -0800

I have a bias against anything hosted or coming out of Germany, and from Western Europe

Why don't you run - say - three Hubs in three areas and have clones on them?
Are you afraid of censorship or that someody copies your content (spy on you)?

FourOh-LLC

Wed, 10 Mar 2021 07:06:52 -0800

I run four hubs, and one is going to be a Secondary for this current realm in a month or so. I use OVH and Linode, for now, but I prefer to use at least three.

Let me put it this way: Walkaway made the mistake to believe that keeping their grassroots on the platform of their political opponent is a good idea - the same happened to Gab years back, and to Parler weeks before. Just how many more lessons you need before you learn from REALITY? Censorship, computer crime is real, if you are not worried you have no survival instincts.

Einer von Vielen

Wed, 10 Mar 2021 07:40:22 -0800

if you are not worried you have no survival instincts.

That's why I run my two small faḿily hubs inside of our flats

Einer von Vielen

Wed, 10 Mar 2021 08:30:52 -0800 last edited: Wed, 10 Mar 2021 08:49:11 -0800

/update/

result of fio for netcup.de
Disk: 320 GB SAS
RAM DDR 4 ECC: 8 GB
CPUs: 2 dediziert
Processor: Intel® Xeon® Gold 6140 / 6230

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randread
test: (g=0): rw=randread, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=64
fio-3.12
Starting 1 process
test: Laying out IO file (1 file / 4096MiB)
Jobs: 1 (f=1): [r(1)][100.0%][r=337MiB/s][r=86.4k IOPS][eta 00m:00s]
test: (groupid=0, jobs=1): err= 0: pid=22789: Wed Mar 10 17:29:54 2021
  read: IOPS=86.5k, BW=338MiB/s (354MB/s)(4096MiB/12126msec)
   bw (  KiB/s): min=319456, max=373325, per=99.92%, avg=345625.87, stdev=13764.03, samples=24
   iops        : min=79864, max=93331, avg=86406.46, stdev=3440.99, samples=24
  cpu          : usr=14.00%, sys=48.10%, ctx=46238, majf=0, minf=71
  IO depths    : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.1%, 16=0.1%, 32=0.1%, >=64=100.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.1%, >=64=0.0%
     issued rwts: total=1048576,0,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=64

Run status group 0 (all jobs):
   READ: bw=338MiB/s (354MB/s), 338MiB/s-338MiB/s (354MB/s-354MB/s), io=4096MiB (4295MB), run=12126-12126msec

Disk stats (read/write):
  sda: ios=1040429/4, merge=0/5, ticks=657859/5, in_queue=647520, util=99.04%

Max Kostikov

Wed, 10 Mar 2021 09:16:54 -0800

@Einer von Vielen (einer@misty.digitalesparadies.de)

result of fio for netcup.de

Wow! It's very fast.
But I thought you're use VPS. Will be interesting is somebody who use VPS on Netcup will perform the same test.
Thank you very much!

FourOh-LLC

Wed, 10 Mar 2021 08:33:09 -0800

Nice job, but what am I looking at? I have no reference for comparison.

Einer von Vielen

Wed, 10 Mar 2021 08:35:28 -0800

Nice job, but what am I looking at? I have no reference for comparison.

@Max Kostikov (kostikov@zotum.net) What is your interpretation of the results?

Einer von Vielen

Wed, 10 Mar 2021 08:42:06 -0800 last edited: Wed, 10 Mar 2021 08:42:36 -0800

I tested the same command on one of my hubs AT HOME, Intel(R) Core(TM) i5-8259U CPU @ 2.30GHz

The output is

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randread
test: (g=0): rw=randread, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=64
fio-3.12
Starting 1 process
test: Laying out IO file (1 file / 4096MiB)
Jobs: 1 (f=1): [r(1)][100.0%][r=1303MiB/s][r=334k IOPS][eta 00m:00s]
test: (groupid=0, jobs=1): err= 0: pid=19161: Wed Mar 10 17:37:51 2021
  read: IOPS=332k, BW=1296MiB/s (1359MB/s)(4096MiB/3160msec)
   bw (  MiB/s): min= 1280, max= 1308, per=100.00%, avg=1297.99, stdev=11.35, samples=6
   iops        : min=327762, max=334892, avg=332284.33, stdev=2904.57, samples=6
  cpu          : usr=29.69%, sys=69.04%, ctx=6080, majf=0, minf=71
  IO depths    : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.1%, 16=0.1%, 32=0.1%, >=64=100.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.1%, >=64=0.0%
     issued rwts: total=1048576,0,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=64

Run status group 0 (all jobs):
   READ: bw=1296MiB/s (1359MB/s), 1296MiB/s-1296MiB/s (1359MB/s-1359MB/s), io=4096MiB (4295MB), run=3160-3160msec

Disk stats (read/write):
  nvme0n1: ios=1025187/37, merge=0/16, ticks=162151/49, in_queue=163240, util=100.00%

FourOh-LLC

Wed, 10 Mar 2021 08:52:12 -0800

What I meant is - this is more than simply a comparison of values with the same unit of measurement. I have no idea how to conclude any performance clues. Its like comparing electrical sources - one is energy, the other is power. Or torque vs. speed, and so forth.

Einer von Vielen

Wed, 10 Mar 2021 09:05:42 -0800

And  this output is from my second private hub AT HOME, a Raspi 4

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randread
test: (g=0): rw=randread, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=64
fio-3.12
Starting 1 process
test: Laying out IO file (1 file / 4096MiB)
Jobs: 1 (f=1): [r(1)][100.0%][r=13.6MiB/s][r=3483 IOPS][eta 00m:00s]
test: (groupid=0, jobs=1): err= 0: pid=30145: Wed Mar 10 18:02:04 2021
  read: IOPS=2816, BW=11.0MiB/s (11.5MB/s)(4096MiB/372234msec)
   bw (  KiB/s): min= 7904, max=14328, per=99.95%, avg=11261.88, stdev=294.46, samples=744
   iops        : min= 1976, max= 3582, avg=2815.43, stdev=73.61, samples=744
  cpu          : usr=2.91%, sys=22.47%, ctx=1122569, majf=0, minf=81
  IO depths    : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.1%, 16=0.1%, 32=0.1%, >=64=100.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.1%, >=64=0.0%
     issued rwts: total=1048576,0,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=64

Run status group 0 (all jobs):
   READ: bw=11.0MiB/s (11.5MB/s), 11.0MiB/s-11.0MiB/s (11.5MB/s-11.5MB/s), io=4096MiB (4295MB), run=372234-372234msec

Disk stats (read/write):
  mmcblk0: ios=1045527/525, merge=3468/270, ticks=23709640/17549, in_queue=23727188, util=100.00%

Einer von Vielen

Wed, 10 Mar 2021 09:09:19 -0800

Wow!!! What differences.

Max Kostikov

Wed, 10 Mar 2021 09:14:25 -0800

@Einer von Vielen (einer@misty.digitalesparadies.de)

Wow!!! What differences.

Yeah, 334k vs 3483 iops is about x100 times faster

Einer von Vielen

Wed, 10 Mar 2021 09:10:20 -0800

It seems to be about read-write perfomance. @Max Kostikov (kostikov@zotum.net) might tell us a little bit.

FourOh-LLC

Wed, 10 Mar 2021 09:47:34 -0800

@Max Kostikov (kostikov@zotum.net) Thanks, this is really useful to know and its pretty easy to do..
Here is the result for the new secondary zot.factoryfouroh.net on OVH.. I have the same VPS for the third year now, a migration was never announced, so it must be aging hardware. I have to make sure the zot.kane-il.us going to have better stats than these :D

test: (g=0): rw=randread, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=64
fio-3.12
Starting 1 process
test: Laying out IO file (1 file / 4096MiB)
Jobs: 1 (f=1): [r(1)][100.0%][r=8124KiB/s][r=2031 IOPS][eta 00m:00s]
test: (groupid=0, jobs=1): err= 0: pid=11688: Wed Mar 10 12:30:43 2021
  read: IOPS=1998, BW=7994KiB/s (8186kB/s)(4096MiB/524696msec)
   bw (  KiB/s): min= 4968, max= 9592, per=100.00%, avg=7993.14, stdev=177.07, samples=1049
   iops        : min= 1242, max= 2398, avg=1998.28, stdev=44.25, samples=1049
  cpu          : usr=1.10%, sys=3.02%, ctx=625318, majf=0, minf=71
  IO depths    : 1=0.1%, 2=0.1%, 4=0.1%, 8=0.1%, 16=0.1%, 32=0.1%, >=64=100.0%
     submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
     complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.1%, >=64=0.0%
     issued rwts: total=1048576,0,0,0 short=0,0,0,0 dropped=0,0,0,0
     latency   : target=0, window=0, percentile=100.00%, depth=64

Run status group 0 (all jobs):
   READ: bw=7994KiB/s (8186kB/s), 7994KiB/s-7994KiB/s (8186kB/s-8186kB/s), io=4096MiB (4295MB), run=524696-524696msec

Disk stats (read/write):
  sda: ios=1048514/269, merge=0/257, ticks=33541783/5897, in_queue=33336000, util=99.38%

Waitman Gobble

Wed, 10 Mar 2021 11:21:41 -0800

I use RamNode, Contabo, Vultr and Linnode. Have not had any bad experiences i can think of. RamNode is cheap. 7 of my RamNode instances randomly drop their ip assignment in Seatle and Atlanta. Other locations are fine. I could never get to the bottom of why this was happening but I made a Perl script to check the interface and set it back up if the network vanishes. That has been working.

Max Kostikov

Wed, 10 Mar 2021 11:25:26 -0800

@Waitman Gobble (wago@zap.dog)

Contabo

Contabo limiting disk performance with 400 iops on their VPS.

Anmol Sharma

Wed, 10 Mar 2021 11:29:35 -0800

@Max Kostikov (kostikov@zotum.net)
Contabo has one plan with hybrid disk and other is with ssd.
One of my client is facing lags and performance issue on hybrid plan.

Max Kostikov

Wed, 10 Mar 2021 11:32:30 -0800

@Anmol Sharma (anmol@anmol.net.in)

Contabo has one plan with hybrid disk and other is with ssd.

I'm talking about SSD plan. 400 iops on SSD.

FourOh-LLC

Wed, 10 Mar 2021 11:32:15 -0800

Vultr has an interesting deal on IPv6-only VPS.. I think its under 3 USD / pay as go. I got one, then I learned that Webmin does not support IPv6 - so I dropped it. For those who can manage from the shell it should be a good deal even with low performance.