Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


13th anniversary promotions - NVMe Compute & RAID10 Storage VMs - Page 8
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

13th anniversary promotions - NVMe Compute & RAID10 Storage VMs

1568101117

Comments

  • @DP said:

    @ajaibdigital said:
    @hosthatch invoice 400665
    Hope will get double bw 😁

    You are more likely to get doubled wait time now :smiley:

    looks like not doubled wait time, but 10x faster than the thread
    its ready to use now 😁

    btw thank's and Happy anniversary @hosthatch !

  • I ordered 2 servers and paid for them but they display this line, is it normal?
    "This server is in a pending state and may be awaiting its invoice to be paid for or is still being built."

  • plumbergplumberg Veteran

    @tonytran said:
    I ordered 2 servers and paid for them but they display this line, is it normal?
    "This server is in a pending state and may be awaiting its invoice to be paid for or is still being built."

    Very normal. From the first post... @hosthatch mentions...

    What is the setup time?
    The setup time is 10 working days (2 weeks).

    Thanked by 1tonytran
  • @plumberg said:

    @tonytran said:
    I ordered 2 servers and paid for them but they display this line, is it normal?
    "This server is in a pending state and may be awaiting its invoice to be paid for or is still being built."

    Very normal. From the first post... @hosthatch mentions...

    What is the setup time?
    The setup time is 10 working days (2 weeks).

    Thanks

  • Fow someone who need yabs in SG

    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2024-04-22                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Wed May  1 01:09:27 WIB 2024
    
    Basic System Information:
    ---------------------------------
    Uptime     : 0 days, 0 hours, 47 minutes
    Processor  : AMD EPYC 7513 32-Core Processor
    CPU cores  : 6 @ 2595.124 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ✔ Enabled
    RAM        : 23.7 GiB
    Swap       : 0.0 KiB
    Disk       : 270.7 GiB
    Distro     : Ubuntu 20.04.3 LTS
    Kernel     : 5.4.0-91-generic
    VM Type    : KVM
    IPv4/IPv6  : ✔ Online / ✔ Online
    
    IPv6 Network Information:
    ---------------------------------
    ISP        : HostHatch
    ASN        : AS63473 HostHatch, LLC
    Host       : HostHatch, LLC
    Location   : Singapore, North West (03)
    Country    : Singapore
    
    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/vda1):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 236.94 MB/s  (59.2k) | 1.85 GB/s    (28.9k)
    Write      | 237.57 MB/s  (59.3k) | 1.86 GB/s    (29.1k)
    Total      | 474.52 MB/s (118.6k) | 3.71 GB/s    (58.0k)
               |                      |                     
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 1.74 GB/s     (3.4k) | 1.83 GB/s     (1.7k)
    Write      | 1.84 GB/s     (3.5k) | 1.95 GB/s     (1.9k)
    Total      | 3.59 GB/s     (7.0k) | 3.78 GB/s     (3.6k)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping           
    -----           | -----                     | ----            | ----            | ----           
    Clouvider       | London, UK (10G)          | 438 Mbits/sec   | 59.6 Mbits/sec  | 355 ms         
    Eranium         | Amsterdam, NL (100G)      | busy            | 71.6 Mbits/sec  | --             
    Telia           | Helsinki, FI (10G)        | busy            | busy            | 396 ms         
    Uztelecom       | Tashkent, UZ (10G)        | 1.12 Gbits/sec  | 1.09 Gbits/sec  | 275 ms         
    Leaseweb        | Singapore, SG (10G)       | 7.56 Gbits/sec  | 6.65 Gbits/sec  | 3.34 ms        
    Clouvider       | Los Angeles, CA, US (10G) | 886 Mbits/sec   | 24.0 Mbits/sec  | 177 ms         
    Leaseweb        | NYC, NY, US (10G)         | busy            | 929 Kbits/sec   | --             
    Edgoo           | Sao Paulo, BR (1G)        | busy            | 285 Mbits/sec   | 384 ms         
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping           
    -----           | -----                     | ----            | ----            | ----           
    Clouvider       | London, UK (10G)          | busy            | 9.05 Mbits/sec  | 354 ms         
    Eranium         | Amsterdam, NL (100G)      | busy            | busy            | --             
    Uztelecom       | Tashkent, UZ (10G)        | 422 Mbits/sec   | 144 Mbits/sec   | 274 ms         
    Leaseweb        | Singapore, SG (10G)       | busy            | 5.62 Gbits/sec  | 2.15 ms        
    Clouvider       | Los Angeles, CA, US (10G) | 808 Mbits/sec   | 922 Mbits/sec   | 179 ms         
    Leaseweb        | NYC, NY, US (10G)         | 483 Mbits/sec   | 514 Mbits/sec   | 294 ms         
    Edgoo           | Sao Paulo, BR (1G)        | busy            | 387 Mbits/sec   | 384 ms         
    
    Geekbench 6 Benchmark Test:
    ---------------------------------
    Test            | Value                         
                    |                               
    Single Core     | 1387                          
    Multi Core      | 5703                          
    Full Test       | https://browser.geekbench.com/v6/cpu/5917084
    
    Thanked by 2sliix vastness4594
  • stxshstxsh Member

    @Daniel15 said:

    @vpn2024 said: But I noticed the offer had different wording, cpu% dedicated vs cores dedicated.

    CPU% dedicated makes more sense, since they're not actually dedicated cores. Dedicated cores would imply usage of CPU pinning / processor affinity to actually dedicate particular cores to VMs (i.e. no other VM can ever use them), which they're not using.

    As far as I know (@hosthatch please correct me if I'm wrong) they're counting threads, not cores. An example of a CPU they use is the AMD EPYC 7513, which has 32 cores / 64 threads. A VM with "1 dedicated core" or "100% dedicated CPU" can use 1/64 of the CPU's power constantly, and the other non-dedicated cores (actually threads) are for burst usage only.

    Yeah that's how I interpreted as well.. from a CPU load average perspective:

    6 CPU cores (2 dedicated, 4 fair-shared cores)

    means, your load average can be 2.0 24/7.. but it can raise up to 6.0 on "fair share" levels/bursts, but not consistently.

    2 CPU cores (0.5 dedicated core, 1.5 fairly shared cores)

    means, your load average can be .5 24/7.. but you can have it raise to 2.0 on occasions.. sound right @hosthatch?

    I can't seem to find it in your AUP, but are there any numbers on what you consider "fair" or does that depend on how active the node is?

  • stxshstxsh Member
    edited April 30

    I just got my first VM activated :smile: .. @hosthatch suggestion for the next version of your panel: it would be nice if you started the billing cycle on the VM provision date and not the purchase date. If a customer pays for a year, and it takes 2 weeks to provision, it should be a year from that date.

    Edit #2: Also, a "paste" feature for console would be appreciated as well.

    Also: cot damn! this panel is pretty.

    Edit: should snapshots return "Sorry, the page you are looking for could not be found." ?

  • DPDP Administrator, The Domain Guy

    Thanked by 2shruub ehab
  • xyzzzxyzzz Member
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2024-04-22                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Tue Apr 30 20:03:19 UTC 2024
    
    Basic System Information:
    ---------------------------------
    Uptime     : 0 days, 1 hours, 24 minutes
    Processor  : AMD EPYC 7R13 48-Core Processor
    CPU cores  : 6 @ 2645.030 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ✔ Enabled
    RAM        : 23.6 GiB
    Swap       : 0.0 KiB
    Disk       : 274.9 GiB
    Distro     : Debian GNU/Linux 12 (bookworm)
    Kernel     : 6.1.0-9-amd64
    VM Type    : KVM
    IPv4/IPv6  : ✔ Online / ❌ Offline
    
    IPv4 Network Information:
    ---------------------------------
    ISP        : HostHatch
    ASN        : AS63473 HostHatch, LLC
    Host       : HostHatch LLC
    Location   : Los Angeles, California (CA)
    Country    : United States
    
    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/vda1):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 278.39 MB/s  (69.5k) | 2.54 GB/s    (39.7k)
    Write      | 279.13 MB/s  (69.7k) | 2.55 GB/s    (39.9k)
    Total      | 557.52 MB/s (139.3k) | 5.10 GB/s    (79.7k)
               |                      |                     
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ---- 
    Read       | 2.55 GB/s     (4.9k) | 2.52 GB/s     (2.4k)
    Write      | 2.68 GB/s     (5.2k) | 2.68 GB/s     (2.6k)
    Total      | 5.24 GB/s    (10.2k) | 5.21 GB/s     (5.0k)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping           
    -----           | -----                     | ----            | ----            | ----           
    Clouvider       | London, UK (10G)          | 1.17 Gbits/sec  | 1.34 Gbits/sec  | 137 ms         
    Eranium         | Amsterdam, NL (100G)      | 1.00 Gbits/sec  | 1.17 Gbits/sec  | 156 ms         
    Telia           | Helsinki, FI (10G)        | busy            | busy            | 150 ms         
    Uztelecom       | Tashkent, UZ (10G)        | 583 Mbits/sec   | 704 Mbits/sec   | 238 ms         
    Leaseweb        | Singapore, SG (10G)       | 865 Mbits/sec   | 1.06 Gbits/sec  | 169 ms         
    Clouvider       | Los Angeles, CA, US (10G) | 9.33 Gbits/sec  | 6.95 Gbits/sec  | 0.420 ms       
    Leaseweb        | NYC, NY, US (10G)         | 1.97 Gbits/sec  | 2.57 Gbits/sec  | 73.5 ms        
    Edgoo           | Sao Paulo, BR (1G)        | 842 Mbits/sec   | 1.03 Gbits/sec  | 180 ms         
    
    Geekbench 6 Benchmark Test:
    ---------------------------------
    Test            | Value                         
                    |                               
    Single Core     | 1621                          
    Multi Core      | 7017                          
    Full Test       | https://browser.geekbench.com/v6/cpu/5918138
    
    YABS completed in 10 min 49 sec
    
    Thanked by 1Smigit
  • Daniel15Daniel15 Veteran
    edited April 30

    @stxsh said: means, your load average can be 2.0 24/7

    You need to look at CPU usage rather than load average. Load average doesn't always correlate with CPU usage.

    You can have a high load average with very low CPU usage, if processes are blocked on I/O. Load average is the number of processes currently running on the CPU (maximum 6 if you have 6 threads), plus the number of processes that would be using the CPU but are in an uninterruptible sleep state, blocked by something (usually I/O, but it could also be CPU time if CPU usage is at 100%).

    You can see this with NFS if you connect to an NFS server but then the server becomes unreachable (e.g. you shut it down). For example, the load average could be a bit over 30 if 30 processes are blocked waiting for files from the NFS share, even though CPU usage might be close to 0%.

    The opposite is true though: High CPU usage will always result in higher load average.

    @stxsh said: : it would be nice if you started the billing cycle on the VM provision date and not the purchase date

    They usually update the billing cycle date, eventually.

    Thanked by 4stxsh emgh xyzzz nick_
  • yongsikleeyongsiklee Member, Patron Provider

    Virtualization supported?

  • anrikazanrikaz Member

    I ordered a Singapore 24GB VPS, but they deployed it in Poland instead. I have opened a support ticket and hope they will resolve the issue with my VPS soon.

  • sliixsliix Member

    @yongsiklee said:
    Virtualization supported?

    Should be based on YABS.

    Thanked by 1yongsiklee
  • yongsikleeyongsiklee Member, Patron Provider

    @sliix said:

    @yongsiklee said:
    Virtualization supported?

    Should be based on YABS.

    Gotcha!

  • aaxaaaaxaa Member

    YABS in Stockholm

    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2024-04-22                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Wed May  1 01:08:15 UTC 2024
    
    Basic System Information:
    ---------------------------------
    Uptime     : 1 days, 1 hours, 8 minutes
    Processor  : AMD EPYC 7542 32-Core Processor
    CPU cores  : 6 @ 2899.998 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ✔ Enabled
    RAM        : 23.7 GiB
    Swap       : 0.0 KiB
    Disk       : 270.6 GiB
    Distro     : Ubuntu 22.04.1 LTS
    Kernel     : 5.15.0-46-generic
    VM Type    : KVM
    IPv4/IPv6  : ✔ Online / ✔ Online
    
    IPv6 Network Information:
    ---------------------------------
    ISP        : HostHatch
    ASN        : AS63473 HostHatch, LLC
    Host       : HostHatch, LLC
    Location   : Stockholm, Stockholm County (AB)
    Country    : Sweden
    
    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/vda1):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 129.73 MB/s  (32.4k) | 1.17 GB/s    (18.3k)
    Write      | 130.07 MB/s  (32.5k) | 1.18 GB/s    (18.4k)
    Total      | 259.80 MB/s  (64.9k) | 2.35 GB/s    (36.8k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 2.20 GB/s     (4.2k) | 2.66 GB/s     (2.6k)
    Write      | 2.31 GB/s     (4.5k) | 2.84 GB/s     (2.7k)
    Total      | 4.51 GB/s     (8.8k) | 5.51 GB/s     (5.3k)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping
    -----           | -----                     | ----            | ----            | ----
    Clouvider       | London, UK (10G)          | 5.90 Gbits/sec  | 2.15 Gbits/sec  | 31.1 ms
    Eranium         | Amsterdam, NL (100G)      | 8.14 Gbits/sec  | 7.05 Gbits/sec  | --
    Telia           | Helsinki, FI (10G)        | busy            | busy            | 7.06 ms
    Uztelecom       | Tashkent, UZ (10G)        | 2.11 Gbits/sec  | 1.43 Gbits/sec  | 101 ms
    Leaseweb        | Singapore, SG (10G)       | 395 Mbits/sec   | 509 Mbits/sec   | 325 ms
    Clouvider       | Los Angeles, CA, US (10G) | 951 Mbits/sec   | 929 Mbits/sec   | 146 ms
    Leaseweb        | NYC, NY, US (10G)         | 1.51 Gbits/sec  | 1.71 Gbits/sec  | 102 ms
    Edgoo           | Sao Paulo, BR (1G)        | 733 Mbits/sec   | 626 Mbits/sec   | 193 ms
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed      | Ping
    -----           | -----                     | ----            | ----            | ----
    Clouvider       | London, UK (10G)          | 5.99 Gbits/sec  | 2.38 Gbits/sec  | 30.5 ms
    Eranium         | Amsterdam, NL (100G)      | busy            | busy            | --
    Uztelecom       | Tashkent, UZ (10G)        | 1.63 Gbits/sec  | 527 Mbits/sec   | 100 ms
    Leaseweb        | Singapore, SG (10G)       | busy            | busy            | 324 ms
    Clouvider       | Los Angeles, CA, US (10G) | 1.03 Gbits/sec  | busy            | 145 ms
    Leaseweb        | NYC, NY, US (10G)         | 1.61 Gbits/sec  | 1.74 Gbits/sec  | --
    Edgoo           | Sao Paulo, BR (1G)        | 744 Mbits/sec   | 776 Mbits/sec   | 191 ms
    
    Geekbench 6 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 1308
    Multi Core      | 5381
    Full Test       | https://browser.geekbench.com/v6/cpu/5920394
    
    YABS completed in 16 min 14 sec
    
    Thanked by 1zrj766
  • sputeksputek Member

    @hosthatch said:

    @iandrewc said:

    @hosthatch said:

    @sputek said:

    @iandrewc said:

    @sputek said:

    @hosthatch said: 6 CPU cores (2 dedicated, 4 fair-shared cores) ROME/MILAN

    How does this actually work? Do we get 2 virtual cores pinned to the host CPU cores (ie no steal time on these cores)?

    Can someone post /proc/stat on previous deals with similar CPU allocation?

    This is the AMD 6 CPU cores (200% dedicated, burstable up to 600%); 16 GB RAM 2023 BF Deal with the 3 year pay bonuses.

    cpu  13306170 50915 3386761 1427241968 84096 0 94953 2208790 0 0
    cpu0 1782275 7561 477223 238699544 10197 0 15124 268099 0 0
    cpu1 2461433 9262 711941 236959773 16919 0 45156 602277 0 0
    cpu2 2539845 9285 648349 237412353 13707 0 3595 382177 0 0
    cpu3 2387271 8060 524828 237930068 9625 0 2581 298552 0 0
    cpu4 2161832 10030 500965 238238733 13456 0 2179 268738 0 0
    cpu5 1973511 6715 523453 238001495 20190 0 26315 388945 0 0
    

    Thank you. I guess you bought https://lowendtalk.com/discussion/190331/black-friday-2023-nvme-and-storage-deals/p1

    All cores have similar amount of CPU steal (3rd column from the end), so it doesn't have dedicated cores. But I noticed the offer had different wording, cpu% dedicated vs cores dedicated.

    @hosthatch Could you clarify? I'd purchase it immediately if the current offer actually has dedicated cores. Getting no steal is very important for latency-sensitive workloads

    If you see steal, you should open a ticket.

    @vpn2024 said: Best effort dedicated

    No. We're not one of those German providers.

    @hosthatch So there should not be steal on the BF 2023 Compute VMs or just not on these 13th anniversary deals?

    There should never be any abnormal steal on any of HostHatch Compute NVMe VM service, promotional or not. By abnormal, I mean ~3% or more.

    OK, so they are not dedicated cores.

    N dedicated cores and N*100% CPU usage allowance are two totally different things. Your offer is still well priced for the latter but please use correct terms. The "6 CPU cores (2 dedicated, 4 fair-shared cores)" can only be read as 2 cores are dedicated which is apparently not true.

    It is rare to combine dedicated and non-dedicated cores so I asked just to be sure... I'm glad I didn't waste my money.

    If you see < 3% steal, we will not go out of our way to fix that. If your application is that sensitive and needs to be making millions, then you should be considering AWS, and not a VPS provider.

    I don't understand your hostility here?

    VPS absolutely work, like those from Hetzner Cloud or Netcup (but in limited locations, hence I was interested in your offer...), as long as CPU cores are actually dedicated and guaranteed not to freeze for 10 milliseconds randomly (ie no steal)

    Not sure why I have to explain this to someone who's in the industry for years tbh

  • emghemgh Member
    edited May 1

    @sputek I’d call you way more rude than @hosthatch, although, sure, his initial tone could have been more professional, but I also appreciate him being himself and honest

    Also, are you comparing against Hetzner’s normal ”dedicated” line? I’m not sure if those cores are pinned, but I don’t believe so

    Also, their 16GB is €24.49/mo and their 32GB is €48.49, let’s say that they’d charge €36.49 if they actually offered a 24GB version

    The 24GB offer is (without the 2-year discount) $9.58/mo (albeit paid yearly). That’s €8.99 or over four times as much per month

    Therefore, it’s not even comparable. I do agree that the marketing around cores sucks (how did we even get to the point where it should just be understood that a core actually is a thread..?) but that goes for almost any provider in existance

    @hosthatch what are the fair use limits and how do you handle ”abuse”, by limiting to the dedicated cores, by notifying the user, or by other means?

    Edit: I regularely open up Spreadsheets and see if for some reason I can create value in this market and start something up. I do offer managed hosting through my company, but I haven’t yet even theoretically been able to make it work selling VPS at LET pricing. This offer is crazy competitive when you calculate the economics of it

  • hosthatchhosthatch Patron Provider, Top Host, Veteran
    edited May 1

    @sputek said: OK, so they are not dedicated cores.

    Yes they are. At any time, you can request for the fair shared cores to be removed from your server, to have only dedicated cores. That is the definition of dedicated CPU. You will see as close to 0 steal as possible once you do that.

    If you think DigitalOcean dedicated CPU cores, which likely cost 10x more, are pinned to your VM, then you have another surprise coming.

    @sputek said: VPS absolutely work, like those from Hetzner Cloud or Netcup (but in limited locations, hence I was interested in your offer...), as long as CPU cores are actually dedicated and guaranteed not to freeze for 10 milliseconds randomly (ie no steal)

    Can you share your steal graph from both of these providers, showing 0.0% steal over a long period of time? I don't even have 0.0% steal on my baremetals, but obviously you seem to know more as you have suggested with the kind "Not sure why I have to explain this to someone who's in the industry for years tbh", so please enlighten me, but with actual data and graphs, thanks.

    The only way you get 0.0% steal on a public cloud VM is by patching your kernels, which we do not, our of respect to transparency. But again, as you have suggested that you know far more than me, so I look forward to being enlightened by your new data.

    Thanked by 3emgh nick_ iandrewc
  • hosthatchhosthatch Patron Provider, Top Host, Veteran

    Nevermind, took a quick look at your profile, you signed up to make these comments to stir this up, so I will let it go and let the discussion go back to the big boys :)

    Thanked by 1iandrewc
  • sliixsliix Member
    edited May 1

    @sputek said: I don't understand your hostility here?

    I don't think @hosthatch being hostile here. Just stating facts. Although surely could've been worded better.

    With that being said, surely we can discuss it further. I would also wanna know more about how "dedicated" is being defined (at least for @hosthatch's services).

    Edit: saw more explanation being given now.

  • sputeksputek Member

    @emgh said:
    @sputek I’d call you way more rude than @hosthatch, although, sure, his initial tone could have been more professional, but I also appreciate him being himself and honest

    @hosthatch I apologize if I sounded rude. That was my intention, but I wanted to confirm I wasn't going to purchase something I can't use.

    Also, are you comparing against Hetzner’s normal ”dedicated” line? I’m not sure if those cores are pinned, but I don’t believe so

    I don't see clock jumps on Hetzner Cloud "dedicated vCPU" line so I assume it is pinned. The cheapest plan costs about €12.49 for 2 dedicated cores. With much less RAM, yes, but I don't need that much so I ignored the difference. Same with bandwidth allocation. Netcup is cheaper, and also doesn't have jumps (caveat is that they're only in Europe).

    I'm not saying HostHatch's offering is bad. I argue however it's incorrectly labelled as "dedicated CPU" when it's not.

    @hosthatch said: Can you share your steal graph from both of these providers, showing 0.0% steal over a long period of time?

    Hetzner Cloud hides steal from VM guests so it is stuck at 0. Anyone can confirm it. Doesn't necessarily mean there is no actual steal, but I don't observe any indication of it.

    Not a current customer of Netcup. Can someone post the first lines of /proc/stat on their dedicated lines?

    I don't even have 0.0% steal on my baremetals, but obviously you seem to know more as you have suggested with the kind "Not sure why I have to explain this to someone who's in the industry for years tbh", so please enlighten me, but with actual data and graphs, thanks.

    What? Indeed, you seem to have no knowledge about how steal works. CPU steal time accounting is done by the VM host. Guest queries the value in the MSR. Naturally, you have no steal on baremetal (nor surprise CPU freeze). Check your /proc/stat.

    https://www.kernel.org/doc/Documentation/virtual/kvm/msr.txt

  • AcacyAcacy Member
    edited May 1

    Singapore Node

    YABS

    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    #              Yet-Another-Bench-Script              #
    #                     v2024-04-22                    #
    # https://github.com/masonr/yet-another-bench-script #
    # ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
    
    Wed May  1 07:31:57 UTC 2024
    
    Basic System Information:
    ---------------------------------
    Uptime     : 0 days, 0 hours, 12 minutes
    Processor  : AMD EPYC 7513 32-Core Processor
    CPU cores  : 4 @ 2595.124 MHz
    AES-NI     : ✔ Enabled
    VM-x/AMD-V : ✔ Enabled
    RAM        : 7.8 GiB
    Swap       : 0.0 KiB
    Disk       : 93.0 GiB
    Distro     : Debian GNU/Linux 11 (bullseye)
    Kernel     : 5.10.0-9-amd64
    VM Type    : KVM
    IPv4/IPv6  : ✔ Online / ✔ Online
    
    IPv6 Network Information:
    ---------------------------------
    ISP        : HostHatch
    ASN        : AS63473 HostHatch, LLC
    Host       : HostHatch, LLC
    Location   : Singapore, North West (03)
    Country    : Singapore
    
    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/vda1):
    ---------------------------------
    Block Size | 4k            (IOPS) | 64k           (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 215.76 MB/s  (53.9k) | 1.85 GB/s    (28.9k)
    Write      | 216.33 MB/s  (54.0k) | 1.86 GB/s    (29.0k)
    Total      | 432.09 MB/s (108.0k) | 3.71 GB/s    (57.9k)
               |                      |
    Block Size | 512k          (IOPS) | 1m            (IOPS)
      ------   | ---            ----  | ----           ----
    Read       | 1.85 GB/s     (3.6k) | 1.85 GB/s     (1.8k)
    Write      | 1.95 GB/s     (3.8k) | 1.98 GB/s     (1.9k)
    Total      | 3.80 GB/s     (7.4k) | 3.84 GB/s     (3.7k)
    
    iperf3 Network Speed Tests (IPv4):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed                                                                                                                   | Ping
    -----           | -----                     | ----            | ----                                                                                                                         | ----
    Clouvider       | London, UK (10G)          | busy            | 502 Mbits/sec                                                                                                                | 314 ms
    Eranium         | Amsterdam, NL (100G)      | 432 Mbits/sec   | 532 Mbits/sec                                                                                                                | --
    Telia           | Helsinki, FI (10G)        | busy            | busy                                                                                                                         | 321 ms
    Uztelecom       | Tashkent, UZ (10G)        | 954 Mbits/sec   | 1.26 Gbits/sec                                                                                                               | 292 ms
    Leaseweb        | Singapore, SG (10G)       | 7.13 Gbits/sec  | 5.91 Gbits/sec                                                                                                               | 2.32 ms
    Clouvider       | Los Angeles, CA, US (10G) | 888 Mbits/sec   | 281 Mbits/sec                                                                                                                | 185 ms
    Leaseweb        | NYC, NY, US (10G)         | busy            | 719 Mbits/sec                                                                                                                | 237 ms
    Edgoo           | Sao Paulo, BR (1G)        | busy            | 212 Mbits/sec                                                                                                                | 387 ms
    
    iperf3 Network Speed Tests (IPv6):
    ---------------------------------
    Provider        | Location (Link)           | Send Speed      | Recv Speed                                                                                                                   | Ping
    -----           | -----                     | ----            | ----                                                                                                                         | ----
    Clouvider       | London, UK (10G)          | busy            | 3.84 Mbits/sec                                                                                                               | 318 ms
    Eranium         | Amsterdam, NL (100G)      | busy            | busy                                                                                                                         | --
    Uztelecom       | Tashkent, UZ (10G)        | busy            | 320 Kbits/sec                                                                                                                | 295 ms
    Leaseweb        | Singapore, SG (10G)       | 7.32 Gbits/sec  | 5.78 Gbits/sec                                                                                                               | 2.04 ms
    Clouvider       | Los Angeles, CA, US (10G) | 769 Mbits/sec   | 784 Mbits/sec                                                                                                                | 184 ms
    Leaseweb        | NYC, NY, US (10G)         | 493 Mbits/sec   | 349 Mbits/sec                                                                                                                | 239 ms
    Edgoo           | Sao Paulo, BR (1G)        | busy            | 347 Mbits/sec                                                                                                                | 387 ms
    
    Geekbench 6 Benchmark Test:
    ---------------------------------
    Test            | Value
                    |
    Single Core     | 1297
    Multi Core      | 4239
    Full Test       | https://browser.geekbench.com/v6/cpu/5923318
    
    YABS completed in 19 min 48 sec
    
  • hosthatchhosthatch Patron Provider, Top Host, Veteran

    @sputek said:

    Great, so not a single graph or actual data then, good argument. You can continue believing whatever you want, and you are welcome to use any of the other providers that are more true in their marketing.

  • There will be Ubuntu 24.04 LTS available any time soon ?

  • webcraftwebcraft Member

    @asderik5633 said:
    There will be Ubuntu 24.04 LTS available any time soon ?

    The entire ISO and Template lineup could be refreshed.. most are over a year old. You can of course just update it or for instance upgrade from Ubuntu 22.04 to 24.04 but it's inconvenient.

  • hosthatchhosthatch Patron Provider, Top Host, Veteran

    @webcraft said: The entire ISO and Template lineup could be refreshed

    Indeed, it's one of those things that is on the todo list. Hopefully by the end of this week.

  • lirrrlirrr Member

    Just got a server, super solid for what I use it for, and the price is unbeatable.

    Thanked by 1hosthatch
  • aaxaaaaxaa Member

    I installed Windows just to test it and saw that the CPU was only 2 instead of 6. Is this a mistake or what is the reason?

  • hosthatchhosthatch Patron Provider, Top Host, Veteran

    @aaxaa said:
    I installed Windows just to test it and saw that the CPU was only 2 instead of 6. Is this a mistake or what is the reason?

    Most likely licensing/Windows limitations.

    Thanked by 1aaxaa
  • haolunhaolun Member

    2 CPU cores (0.5 dedicated core, 1.5 fairly shared cores) ROME/MILAN
    6 GB RAM
    75 GB NVMe storage
    10 TB premium bandwidth
    $85.00 / Triennially

Sign In or Register to comment.