Want to play with a REAL cloud VPS?

Gary

http://www.togglebox.com - discount code reddit99 gets you 99% off your first month. Worth a play to see how the other half live?

MrAndroid

There's no such thing as a "Real" Cloud VPS.

japon

Ah, the company that puts London on Paris

Gary

It's more of a cloud VPS than most providers claim to be.

And they misspelled Australia too. Not the best at Geography, but I'm currently unixbenching a 32 core, 32GB, 500GB, gbit instance that cost me $6 for the month. I actually might even make use of it, since I might have some video encoding to do for a client, and this thing should outperform my quad cores at home...

Damian

It's not very flexible. I was hoping for a 32-core instance, with 256mb of ram and 5gb of disk space, and therefore cheap, but the minimum amount of ram I can get with 32 cores is 32gb. And I don't need to pay for that much.

Regardless, for $6.70, why not? I signed up, tried to create a new "virtual machine" but was only allowed to create one with 24 CPU cores.

vmhosts

Be careful with adding CPU cores just because you can. You give the CPU scheduler a lot more work to do and this can dramatically increase CPU ready times.

Adding CPU's when they are not actually needed can actually lower performance. Always scale out with virtual machines not up (if your application can scale that way)

Damian

Even that didn't work:

http://onyx.ipxcore.com/weirderror.png

I've got a support ticket in on it.

vmhosts

Looks like Onapp, if you go to configure a Cloud VM on their ordering page it only allows 4 cores. Maybe the screen you are on here isnt ment to be viewable?

Gary

Yeah, that annoyed me too. I'd have been happy with 32 cores available, but way less drivespace and ram. It'll do for playing with I guess.

`Benchmark Run: Tue Jan 10 2012 18:09:10 - 18:37:30
24 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 23432909.3 lps (10.0 s, 7 samples)
Double-Precision Whetstone 3050.2 MWIPS (9.9 s, 7 samples)
Execl Throughput 1282.2 lps (29.8 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 217570.9 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 57223.1 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 609500.5 KBps (30.1 s, 2 samples)
Pipe Throughput 351108.2 lps (10.0 s, 7 samples)
Pipe-based Context Switching 69183.5 lps (10.0 s, 7 samples)
Process Creation 2556.0 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 3978.9 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1720.3 lpm (60.1 s, 2 samples)
System Call Overhead 424940.5 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 23432909.3 2008.0
Double-Precision Whetstone 55.0 3050.2 554.6
Execl Throughput 43.0 1282.2 298.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 217570.9 549.4
File Copy 256 bufsize 500 maxblocks 1655.0 57223.1 345.8
File Copy 4096 bufsize 8000 maxblocks 5800.0 609500.5 1050.9
Pipe Throughput 12440.0 351108.2 282.2
Pipe-based Context Switching 4000.0 69183.5 173.0
Process Creation 126.0 2556.0 202.9
Shell Scripts (1 concurrent) 42.4 3978.9 938.4
Shell Scripts (8 concurrent) 6.0 1720.3 2867.1
System Call Overhead 15000.0 424940.5 283.3
========
System Benchmarks Index Score 530.8`

Very high results on a few tests, but nothing to shout about in terms of the index score.

dd results are average-bad too.

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, 23.5096 s, 45.7 MB/s

Mind you, with 32GB of ram to mess around with, I'd be better doing everything in /dev/shm/

root@testvps1:/dev/shm# dd if=/dev/zero of=test bs=64k count=250k conv=fdatasync 256000+0 records in 256000+0 records out 16777216000 bytes (17 GB) copied, 9.46868 s, 1.8 GB/s

I know where the scratch drive will be when I'm encoding anything...

Network performance is average at best, crap mostly. Given that this is supposed to be a gbit pipe...

Cachefly:
2012-01-10 18:51:06 (5.05 MB/s) - â/dev/nullâ

Softlayer's Seattle test file
2012-01-10 18:54:41 (4.85 MB/s) - â/dev/nullâ

Linode's Dallas test node
2012-01-10 18:54:58 (86.7 MB/s) - â/dev/nullâ

I have 24 of these, btw.

processor : 23 vendor_id : GenuineIntel cpu family : 6 model : 44 model name : Intel(R) Xeon(R) CPU E5645 @ 2.40GHz stepping : 2 cpu MHz : 2400.084 cache size : 12288 KB fpu : yes fpu_exception : yes cpuid level : 11 wp : yes flags : fpu de tsc msr pae cx8 sep cmov pat clflush mmx fxsr sse sse2 ss ht syscall nx lm constant_tsc rep_good nonstop_tsc pni ssse3 cx16 sse4_1 sse4_2 popcnt hypervisor lahf_lm bogomips : 4800.16 clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management:

So the nodes only have two of these, I guess. I could have 4 VMs each with 8 CPUs, or a combination of that up to 32 cores in total, but no one VM with 32 cores.

vmhosts

whoops sorry my mistake, if you change the cloud to medium or large you can get more cores

vmhosts

@gary said: So the nodes only have two of these, I guess. I could have 4 VMs each with 8 CPUs, or a combination of that up to 32 cores in total, but no one VM with 32 cores.

I still dont think this is a good long term solution. Would be good if you were the only person on that host but as soon as more VM's are sharing the cores, the CPU scheduler will be working hard and wait times will increase with this number of cores assigned to a single virtual machine. This could have some unexpected results for any applications running on the VM

If they are using shared storage for their cloud then it is probably optimised for random IO, hence the average dd results

drmike

@japon said: Ah, the company that puts London an Paris

New York looks to be in Canada.

Los Angeles is inland.

At least their chat popup stays dismissed when closed out. Why can;t the others providers who use it fix theirs?

Damian

Hi,

We've received quite a few orders for large plans with 32GB of RAM just in the past few hours, this is highly suspect. The plan you chose is the same as all these other people, I'm not terminating your account right now because it appears to be a valid domain with website so it may be coincidence. We went from 200GB RAM capacity to only 20GB because of all these orders that we are weeding through and are likely a scam/fraud.

Anyhow, for cores the max VM size is 24 cores. The cloud can be purchased larger because the idea is to create many VMs, ie. buy 32 cores and create 4 8-core VMs.

CPU priority is % of 1 core, ie. you'd need 2400% CPU priority for 24 cores guaranteed. With 100% (1 core guaranteed) and 24 cores the high CPU priority you can set is 4% (4% of 24 cores = 1 full core guaranteed).

Thank you,
Matt
ToggleBox

Looks like they don't want business anyway.

I ended up cancelling. Another LET user pointed out to me that I can get a 24-core VPS account for €2 per month from EDIS.

Gary

How is them making an insane discount code public, and people making use of it, a scam or a fraud? What did they think would happen?

I got en email asking for ID from them, because "Your order failed to pass our Anti-Fraud filters and requires additional verification to be set up."

My account is still active though, and my VM is working.

Damian

They promptly refunded my payment. So nothing to lose, I guess.

djvdorp

@Damian4IPXcore said: I ended up cancelling. Another LET user pointed out to me that I can get a 24-core VPS account for €2 per month from EDIS.

Do you have a link for this? Might be interested too. And specs ofcourse

yomero

@djvdorp said: @Damian4IPXcore said: I ended up cancelling. Another LET user pointed out to me that I can get a 24-core VPS account for €2 per month from EDIS.

Do you have a link for this? Might be interested too. And specs ofcourse

+1

It seems just Chicago? And just VRS? Or the KVM ones have 24 cores too? :P

Damian

http://en.edis.at/vrs-micro_59_82.htm is the one I was looking at.

Shared Dual Intel E5620 (16 cores @2.4 GHz) available in Austria, Switzerland and UK
Shared Dual Intel E5649 (24 cores @2.53 GHz) available in Italy, Germany, U.S.A.

Their USA one is Chicago, as yomero pointed out.

Aidy

Hmm, somewhat tempting to spin one of these up to deal with my QualityServers migration...

Damian

File this under LOL:

Dear Damian,

This is a billing notice that your Invoice no. 01461 which was generated on 01/10/2012 is now overdue.

Your payment method is: VISA / Mastercard / AMEX

Invoice: 01461
Amount Due: $679.60 USD
Due Date: 01/10/2012.

You can login to your client area to view and pay the invoice at >http://manage.togglebox.com/?cmd=clientarea&action=invoices

Thank you,
ToggleBox
http://www.ToggleBox.com

Gary

Billing fail...