Howdy, Stranger!

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


A quick review & benchmark of massiveGrid (promo) VPS
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.

A quick review & benchmark of massiveGrid (promo) VPS

jsgjsg Member, Resident Benchmarker
edited October 19 in Reviews

So, what do you really get from @MassiveGRID ?

Here you go, based on a bit over 10 runs:
First, as usual, processor and memory

Machine: amd64, Arch.: amd64, Model: Intel Core Processor (Broadwell, no TSX)
OS, version: FreeBSD 14.0, Mem.: 1.991 GB
CPU - Cores: 1, Family/Model/Stepping: 6/61/2
Cache: 32K/32K L1d/L1i, 2M L2, 16M L3
Std. Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat
          pse36 cflsh mmx fxsr sse sse2 sse3 pclmulqdq ssse3 fma cx16 pcid
          sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline aes xsave osxsave avx
          f16c rdrnd hypervisor
Ext. Flags: fsgsbase bmi1 avx2 smep bmi2 erms invpcid rdseed adx smap syscall nx
          rdtscp lm lahf_lm lzcnt

AES? Yes
Nested Virt.? No
HW RNG? Yes

ProcMem SC [MB/s]: avg 136.5 - min 44.8 (32.8 %), max 227.0 (166.3 %)
ProcMem MA [MB/s]: avg 215.8 - min 205.5 (95.2 %), max 229.1 (106.2 %)
ProcMem MB [MB/s]: avg 213.6 - min 198.9 (93.1 %), max 233.9 (109.5 %)
ProcMem AES [MB/s]: avg 513.4 - min 497.8 (97.0 %), max 524.7 (102.2 %)
ProcMem RSA [kp/s]: avg 56.4 - min 44.8 (79.4 %), max 63.5 (112.5 %)

First, please note that this is the 2 GB mem. version! As my tests (like generally most programs) profit a lot from more memory, your performance on a 1 GB VPS likely is somewhat lower.

That said, I like that we get AES and rng hardware support and honestly do not care about not having nested virtualization.

I'm also OK wrt the performance. Sure it's not high plus the spread is quite high but it certainly is acceptable for the super-low price and good enough for small tasks like an NS server or even a simple, preferably static, website.

Nothing to celebrate but neither to complain about.

Now, the disk ...

--- Disk 4 KB - Buffered ---
Write seq. [MB/s]: avg 0.81 - min 0.47 (57.7%), max 1.07 (131.4%)
Write rnd. [MB/s]: avg 0.55 - min 0.35 (63.4%), max 0.75 (135.9%)
Read seq. [MB/s]:  avg 5.06 - min 4.38 (86.5%), max 5.74 (113.3%)
Read rnd. [MB/s]:  avg 4.25 - min 3.73 (87.7%), max 5.12 (120.4%)
--- Disk 4 KB - Sync/Direct ---
Write seq. [MB/s]: avg 0.98 - min 0.56 (57.1%), max 1.26 (128.6%)
Write rnd. [MB/s]: avg 0.73 - min 0.45 (62.0%), max 1.31 (180.6%)
Read seq. [MB/s]:  avg 5.01 - min 3.80 (75.8%), max 6.07 (121.0%)
Read rnd. [MB/s]:  avg 4.32 - min 3.64 (84.2%), max 4.94 (114.3%)

--- Disk 64 KB - Buffered ---
Write seq. [MB/s]: avg 8.02 - min 5.73 (71.4%), max 10.20 (127.1%)
Write rnd. [MB/s]: avg 5.97 - min 4.19 (70.2%), max 8.46 (141.7%)
Read seq. [MB/s]:  avg 439.24 - min 360.38 (82.0%), max 488.85 (111.3%)
Read rnd. [MB/s]:  avg 48.46 - min 39.89 (82.3%), max 60.27 (124.4%)
--- Disk 64 KB - Sync/Direct ---
Write seq. [MB/s]: avg 4.18 - min 3.41 (81.6%), max 4.73 (113.2%)
Write rnd. [MB/s]: avg 2.68 - min 2.37 (88.4%), max 3.03 (113.0%)
Read seq. [MB/s]:  avg 424.70 - min 334.82 (78.8%), max 506.10 (119.2%)
Read rnd. [MB/s]:  avg 48.72 - min 37.32 (76.6%), max 58.17 (119.4%)

--- Disk 1 MB - Buffered ---
Write seq. [MB/s]: avg 9.43 - min 6.36 (67.5%), max 11.71 (124.2%)
Write rnd. [MB/s]: avg 29.17 - min 25.27 (86.6%), max 30.46 (104.4%)
Read seq. [MB/s]:  avg 314.94 - min 221.89 (70.5%), max 391.61 (124.3%)
Read rnd. [MB/s]:  avg 219.78 - min 179.20 (81.5%), max 249.01 (113.3%)
--- Disk 1 MB - Sync/Direct ---
Write seq. [MB/s]: avg 8.65 - min 7.43 (85.9%), max 9.67 (111.8%)
Write rnd. [MB/s]: avg 21.68 - min 18.39 (84.8%), max 23.51 (108.5%)
Read seq. [MB/s]:  avg 318.48 - min 242.90 (76.3%), max 375.94 (118.0%)
Read rnd. [MB/s]:  avg 254.12 - min 197.20 (77.6%), max 301.21 (118.5%)
--- Disk IOps (Sync/Direct) ---
Write seq. [MB/s]: avg 0.82 - min 0.66 (80.1%), max 0.98 (118.9%)
IOps             : avg 210.74 - min 168.66 (80.0%), max 250.57 (118.9%)

Pardon me? Not even 1 MB/s (write) IO and just a bit over 200 IOps? What a shit show! Hell, even the (supposed to not be particularly fast) spinning rust storage drive on my @host_c VPS does better! And this one here is sold as "SSD CEPH Storage" (quoted from OP). Yuck!

Finally, connectivity ...

     
--- Network ---

NO OSL ftp.uninett.no [F: 0]
    DL [Mb/s]:      avg 131.3 - min 52.0 (39.6%), max 232.4 (177.0%)
    Ping [ms]:      avg 30.9 - min 27.7 (89.6%), max 37.2 (120.3%)
    Web ping [ms]:  avg 42.8 - min 37.2 (87.0%), max 50.8 (118.8%)

UK LON lon.speedtest.clouvider.net [F: 0]
    DL [Mb/s]:      avg 268.1 - min 199.9 (74.5%), max 351.9 (131.2%)
    Ping [ms]:      avg 2.9 - min 2.8 (96.9%), max 3.1 (107.2%)
    Web ping [ms]:  avg 5.4 - min 2.8 (52.3%), max 16.5 (308.1%)

NL AMS ftp2.nl.freebsd.org [F: 0]
    DL [Mb/s]:      avg 126.9 - min 81.2 (64.0%), max 178.3 (140.6%)
    Ping [ms]:      avg 8.5 - min 8.3 (97.5%), max 9.6 (112.8%)
    Web ping [ms]:  avg 67.8 - min 12.0 (17.7%), max 222.7 (328.3%)

DE FRA fra.lg.core-backbone.com [F: 0]
    DL [Mb/s]:      avg 225.7 - min 90.6 (40.1%), max 327.4 (145.0%)
    Ping [ms]:      avg 14.9 - min 14.7 (98.8%), max 15.0 (100.8%)
    Web ping [ms]:  avg 20.1 - min 14.9 (74.2%), max 56.8 (283.0%)

FR PAR ftp.fr.freebsd.org [F: 0]
    DL [Mb/s]:      avg 69.1 - min 57.4 (83.1%), max 90.2 (130.6%)
    Ping [ms]:      avg 9.7 - min 9.6 (98.7%), max 10.0 (102.8%)
    Web ping [ms]:  avg 110.3 - min 9.8 (8.9%), max 216.9 (196.7%)

IT ROM giano.com.dist.unige.it [F: 1]
    DL [Mb/s]:      avg 145.2 - min 0.0 (0.0%), max 247.3 (170.4%)
    Ping [ms]:      avg 33.1 - min 0.0 (0.0%), max 38.8 (117.3%)
    Web ping [ms]:  avg 45.8 - min 0.0 (0.0%), max 69.5 (151.7%)

RO BUC mirrors.xservers.ro [F: 0]
    DL [Mb/s]:      avg 86.7 - min 39.3 (45.4%), max 140.7 (162.3%)
    Ping [ms]:      avg 43.0 - min 42.4 (98.6%), max 43.6 (101.4%)
    Web ping [ms]:  avg 51.0 - min 42.8 (83.9%), max 84.5 (165.7%)

TR ANK ftp.linux.org.tr - complete fail

RU MOS speedtest.hostkey.ru [F: 1]
    DL [Mb/s]:      avg 100.5 - min 0.0 (0.0%), max 227.8 (226.6%)
    Ping [ms]:      avg 45.7 - min 45.6 (99.7%), max 45.9 (100.4%)
    Web ping [ms]:  avg 60.4 - min 45.9 (76.0%), max 64.2 (106.3%)

RU SIB mirror.truenetwork.ru [F: 0]
    DL [Mb/s]:      avg 61.7 - min 29.4 (47.6%), max 107.7 (174.6%)
    Ping [ms]:      avg 108.9 - min 104.9 (96.4%), max 111.1 (102.0%)
    Web ping [ms]:  avg 183.4 - min 126.6 (69.0%), max 250.6 (136.7%)

IN MUM mirrors.piconets.webwerks.in [F: 7]
    DL [Mb/s]:      avg 19.1 - min 0.0 (0.0%), max 82.9 (434.9%)
    Ping [ms]:      avg 139.5 - min 138.5 (99.3%), max 142.5 (102.1%)
    Web ping [ms]:  avg 195.7 - min 138.6 (70.8%), max 503.9 (257.5%)

SG SGP download.nus.edu.sg [F: 11]
    DL [Mb/s]:      avg 0.0 - min 0.0 (0.0%), max 0.0 (0.0%)
    Ping [ms]:      avg 0.0 - min 0.0 (-nan%), max 0.0 (-nan%)
    Web ping [ms]:  avg 0.0 - min 0.0 (-nan%), max 0.0 (-nan%)

CN ZHJ mirrors.163.com [F: 9]
    DL [Mb/s]:      avg 5.2 - min 0.0 (0.0%), max 34.9 (667.5%)
    Ping [ms]:      avg 242.7 - min 241.3 (99.4%), max 244.8 (100.9%)
    Web ping [ms]:  avg 298.3 - min 241.3 (80.9%), max 610.2 (204.6%)

JP TOK speedtest.tokyo2.linode.com [F: 0]
    DL [Mb/s]:      avg 48.0 - min 42.0 (87.5%), max 49.7 (103.6%)
    Ping [ms]:      avg 225.9 - min 218.7 (96.8%), max 236.0 (104.5%)
    Web ping [ms]:  avg 228.7 - min 222.5 (97.3%), max 236.0 (103.2%)

AU AAR ftp.au.freebsd.org [F: 1]
    DL [Mb/s]:      avg 31.2 - min 0.0 (0.0%), max 38.2 (122.4%)
    Ping [ms]:      avg 291.3 - min 289.1 (99.2%), max 296.3 (101.7%)
    Web ping [ms]:  avg 306.5 - min 290.5 (94.8%), max 378.0 (123.3%)

US MAS ftp5.freebsd.org  - complete fail

US BUF ftp11.freebsd.org - complete fail

US LAX la.speedtest.clouvider.net [F: 0]
    DL [Mb/s]:      avg 40.8 - min 20.2 (49.7%), max 58.2 (142.8%)
    Ping [ms]:      avg 131.6 - min 131.5 (99.9%), max 131.9 (100.2%)
    Web ping [ms]:  avg 139.0 - min 131.5 (94.6%), max 160.9 (115.7%)

BR SAO ftp3.br.freebsd.org [F: 8]
    DL [Mb/s]:      avg 8.1 - min 0.0 (0.0%), max 36.9 (456.2%)
    Ping [ms]:      avg 209.4 - min 207.3 (99.0%), max 212.2 (101.4%)
    Web ping [ms]:  avg 221.9 - min 207.3 (93.4%), max 265.6 (119.7%)

ZA JOB ftp2.za.freebsd.org  - complete fail

KE NAI mirror.liquidtelecom.com [F: 2]
    DL [Mb/s]:      avg 36.7 - min 0.0 (0.0%), max 66.5 (181.4%)
    Ping [ms]:      avg 129.9 - min 129.8 (99.9%), max 130.1 (100.1%)
    Web ping [ms]:  avg 135.7 - min 129.9 (95.7%), max 164.8 (121.4%)

Meh, not great, mainly in terms of the failed US targets, but neither crappy, in particular if you use your VPS mainly within your continent (my VPS is in UK London). Within my VPS's continent, let's call it Europe plus western Asia connectivity actually is quite OK, not great but OK. I mean, within Europe I largely get 100+ Mb/s and hey, I even get some connectivity to Asia and even Ozzyland.

Side note: You also get a very generous amount of traffic (20 TB/mo) for the super-low price.

My conclusion:

First, please keep in mind that these numbers represent a real world load, not just some funny number. Writing to and reading from the disk are real world numbers achieved by real world code (as used in thousands of programs) and not some "let's see how fast that disk is (in theory)", and the "DL" (download) numbers are not some fairy tale numbers based on protocols not used when e.g. your browser downloads a file, and also not numbers with targets that are optimized for high numbers for marketing, but actual download-via-http results from actual http servers e.g. for downloading linux or BSD. Because that is what I - and quite likely you as well -am interested in, not marketing BS or numbers porn.

Those VPSs actually are a pretty good deal - if used for a limited set of jobs. The vCPU (as well as the XEON it's based on) is decent, especially for the super-low price, connectivity is kind of limited but OK within the "greater region", but the disk is just crappy (but reliable if they really are redundant).
In other words: If you're mainly addressing customers/visitors/users within your continent and don't demand too much of it, I think you'll be fine - but IMO do not even dream about putting a database on it or streaming from it!

So, is it a good deal? I'll borrow a statement from @Falzo to summarize it (as I remember it, not word by word) "an IP4 for 50 cents a month? I'll take it". So, following that (largely correct) thinking you even get a halfway decent vCPU and a reasonable amount of memory on top of the cheap IP4, and sadly a mediocre (to put it politely) disk as well.

I personally am happy and think I've made a good deal. But then I know what I'm doing and I use it for a well matching job, and more importantly, a job with minimal disk load (and with my 2 GB version I can afford that ...

Extra "bonus" points for a quite rich selection of OSs and a quite decent panel that, unlike quite a few others, offer all the "knob and buttons", settings, and info one needs.

Big fat minus: the promo thread shit show with 90% bot-like fluff.

TL;DR: a no-brainer for the low price and actually not bad (modulo the crappy disk).

«13

Comments

  • JasonMJasonM Member
    edited October 20

    those 1GB VPS and other too seem oversold.
    one of my friend bought 2GB VPS and he can't even load WordPress Admin area because of 1 MB/s IO and bad network he was complaining... it seems massivegrid has old servers with old hard disk instead of SSD.
    good only for hosting private VPN and nothing else serious like production sites.

    Plus they claim they are 21 years in business and still there are not even 10 reviews or any independent bloggers posting about their vps/dedicated server usage. Google massivegrid and you'll find only information about them and no reviews since last 21 years or since 2014 (when it was started to be called Massivegrid)

    Also I found they sell Cloud Dedicated server with vCores LOL instead of pure Dedicated Cores. The meaning itself becomes useless when you see its just a virtual core (VDS) hosting and not REAL Dedicated Servers. I found that 64GB vCore so callled Cloud Dedicated will cost me $1000+ per month with massivegrid. Rather a user will go with pure bare metal dedicated server for like $150-$250 price range.

    In short, I was inclined to buy their VPS offer (buy for 3 years, get 4th year free) but it seems not useful for my use-case of hosting even a lite-weight/static production site on their so-called HA cloud vps line.

  • zhuyezhuye Member

    I can't believe it. 6 dollars a year for server usage.
    If I'm a businessman, I can't find a place to make a profit; high labor costs, network costs, hardware depreciation and power costs scare me.
    So, I think it's going to be a scam.
    If not, it's network congestion, poor hardware and long response times.

  • MaouniqueMaounique Host Rep, Veteran

    @jsg said: So, what do you really get from @MassiveGRID ?

    Thanks for the benchmark, I was curious myself and ready to buy one.
    I only use these for VPNs, so the traffic part is relevant for me, i.e. the amount (supergenerous at even regular price) and the quality. Speed is not important, but the ping is.

    From what I see, this statement:

    @zhuye said: it's network congestion

    Is not reflecting the reality as the jitter is pretty low during the test, but are you monitoring longtime for disconnects and higher latency? If so, do they happen?

    Thanks :)

    Thanked by 1jsg
  • FalzoFalzo Member

    It's a gamble. And of course depends on the use case. VPN or tunneling the IP somewhere close by etc. can be a win considering that addirional IPs usually are >1.50 per month already everywhere...

    Hosting production stuff on it? Who in their right mind would do that? People who put their valuable prod data on random lowend hosts at cheapest rates can't be helped anyway.

    If this is just a blowup in terms of numbers and potential exit game - so be it. $28 for some fun and a medium chance to have some cheap IP for 4 years 🤷🏻‍♂️

    Thanked by 3jsg emgh dev077
  • jsgjsg Member, Resident Benchmarker

    @JasonM said:
    those 1GB VPS and other too seem oversold.
    one of my friend bought 2GB VPS and he can't even load WordPress Admin area because of 1 MB/s IO and bad network he was complaining... it seems massivegrid has old servers with old hard disk instead of SSD.
    good only for hosting private VPN and nothing else serious like production sites.

    ACK. And I didn't quite clearly hint at that without good reasons. Even with very cheaply purchased 2nd hand equipment their prices (in this promo) IMO just are not doable, let alone tenable, without quite massive (pun intended) overbooking. I just expressed it differently ("high spread", "small tasks", ...) because in a review what I say should be based in data (not merely opinion).

    Plus they claim they are 21 years in business and still there are not even 10 reviews or any independent bloggers posting about their vps/dedicated server usage. Google massivegrid and you'll find only information about them and no reviews since last 21 years or since 2014 (when it was started to be called Massivegrid)

    There I'm more kind of forgiving. What seems to be their normal clientele tends to not write reviews. But yes, it makes one think. My guess is that either their normal model/market didn't work out as planned -or- they have a wish (or IMO more likely a need) to enter other markets, in particular the LE market.

    Plus: Why should one care? Besides potentially going belly up - which can happen even to very well know, often reviewed, and well reputed providers (NexusBytes being an example) - all or at least most I care about is price and bang per buck.

    In short, I was inclined to buy their VPS offer (buy for 3 years, get 4th year free) but it seems not useful for my use-case of hosting even a lite-weight/static production site on their so-called HA cloud vps line.

    Frankly, for me the major driving point was exactly that, the 3 (or 4 years). Having more than a few servers that allows for less worries (like "did I really not miss any payment?"). And at that price it was a no-brainer and I do have tasks I like to know (well, actually presume) to be taken care of for a long while.

    Thanked by 1sucre13
  • jsgjsg Member, Resident Benchmarker

    @Maounique said:

    @jsg said: So, what do you really get from @MassiveGRID ?

    Thanks for the benchmark, I was curious myself and ready to buy one.
    I only use these for VPNs, so the traffic part is relevant for me, i.e. the amount (supergenerous at even regular price) and the quality. Speed is not important, but the ping is.

    Yes, for a VPN the box should do fine, well within the large region, not globally.

    From what I see, this statement:

    @zhuye said: it's network congestion

    Is not reflecting the reality as the jitter is pretty low during the test, but are you monitoring longtime for disconnects and higher latency? If so, do they happen?

    Not yet, besides a benchmark run from time to time, but I'm thinking about automating that, albeit at a somewhat lower frequency (like weekly).

    @Falzo said:
    It's a gamble. And of course depends on the use case. VPN or tunneling the IP somewhere close by etc. can be a win considering that addirional IPs usually are >1.50 per month already everywhere...

    Hosting production stuff on it? Who in their right mind would do that? People who put their valuable prod data on random lowend hosts at cheapest rates can't be helped anyway.

    If this is just a blowup in terms of numbers and potential exit game - so be it. $28 for some fun and a medium chance to have some cheap IP for 4 years 🤷🏻‍♂️

    Yep, that's a good summary I guess. At the end of the day, yes, it is a bet (for us. For the provider it's a good chance) and at a price I'm willing to pay - or loose, quite soon worst case. But I don't think it's a planned exit scam.

  • emghemgh Member, BF Ambassador

    @Falzo said: People who put their valuable prod data on random lowend hosts at cheapest rates can't be helped anyway.

    I mean, sure, this offer is very low-end, but I'm not sure I'd call the host low-end. Their standard prices aren't low at all and this is their first LET offer.

    Only major complaint I have is that I suspect they use the same HW for regular offers, and imagine paying their regular prices and not getting better performance...

  • NanjaNanja Member

    @emgh said:
    Only major complaint I have is that I suspect they use the same HW for regular offers, and imagine paying their regular prices and not getting better performance...

    This is a monthly plan from the MassiveGRID that I got just now.
    16 Cores | 48GB RAM | 512GB Storage

    This is the YABS I just ran.

    root@debian:~# curl -sL yabs.sh | bash

    ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##

    Yet-Another-Bench-Script

    v2024-06-09

    https://github.com/masonr/yet-another-bench-script

    ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ##

    Sun Oct 20 01:15:55 PM UTC 2024

    Basic System Information:

    Uptime : 0 days, 0 hours, 2 minutes
    Processor : Intel(R) Xeon(R) Gold 6130 CPU @ 2.10GHz
    CPU cores : 16 @ 2095.044 MHz
    AES-NI : ✔ Enabled
    VM-x/AMD-V : ✔ Enabled
    RAM : 47.0 GiB
    Swap : 4.0 GiB
    Disk : 498.9 GiB
    Distro : Debian GNU/Linux 12 (bookworm)
    Kernel : 6.1.0-26-amd64
    VM Type : KVM
    IPv4/IPv6 : ✔ Online / ❌ Offline

    IPv4 Network Information:

    ISP : Massivegrid LTD
    ASN : AS49683 MASSIVEGRID LTD
    Host : MassiveGRID
    Location : New York, New York (NY)
    Country : United States

    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/sda3):

    Block Size 4k (IOPS) 64k (IOPS)
    Read 3.98 MB/s (996) 63.08 MB/s (985)
    Write 4.01 MB/s (1.0k) 63.50 MB/s (992)
    Total 7.99 MB/s (1.9k) 126.58 MB/s (1.9k)
    Block Size 512k (IOPS) 1m (IOPS)
    ------ --- ---- ---- ----
    Read 147.08 MB/s (287) 144.54 MB/s (141)
    Write 154.90 MB/s (302) 154.17 MB/s (150)
    Total 301.98 MB/s (589) 298.71 MB/s (291)

    iperf3 Network Speed Tests (IPv4):

    Provider Location (Link) Send Speed Recv Speed Ping
    Clouvider London, UK (10G) 526 Mbits/sec busy 71.8 ms
    Eranium Amsterdam, NL (100G) 587 Mbits/sec 732 Mbits/sec 75.6 ms
    Uztelecom Tashkent, UZ (10G) 400 Mbits/sec 387 Mbits/sec 159 ms
    Leaseweb Singapore, SG (10G) 248 Mbits/sec 376 Mbits/sec 248 ms
    Clouvider Los Angeles, CA, US (10G) 658 Mbits/sec 857 Mbits/sec 67.3 ms
    Leaseweb NYC, NY, US (10G) 1.08 Gbits/sec 919 Mbits/sec 2.18 ms
    Edgoo Sao Paulo, BR (1G) 462 Mbits/sec 515 Mbits/sec 126 ms

    Geekbench 6 Benchmark Test:

    Test | Value
    |
    Single Core | 649
    Multi Core | 3524
    Full Test | https://browser.geekbench.com/v6/cpu/8401136

    YABS completed in 16 min 29 sec

  • emghemgh Member, BF Ambassador

    @Nanja said: This

    $24.16?

  • NanjaNanja Member

    @emgh said:

    @Nanja said: This

    $24.16?

    yes

    Thanked by 1emgh
  • @Nanja said:
    fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/sda3):
    Block Size 4k (IOPS) 64k (IOPS)
    Read 3.98 MB/s (996) 63.08 MB/s (985)
    Write 4.01 MB/s (1.0k) 63.50 MB/s (992)
    Total 7.99 MB/s (1.9k) 126.58 MB/s (1.9k)

    Block Size 512k (IOPS) 1m (IOPS)
    ------ --- ---- ---- ----
    Read 147.08 MB/s (287) 144.54 MB/s (141)
    Write 154.90 MB/s (302) 154.17 MB/s (150)
    Total 301.98 MB/s (589) 298.71 MB/s (291)

    My VPS (2 cores 4 GB RAM located in Frankfurt) produced pretty much the same results on the disk test. Could it be related to the Equinix resources provided to massivegrid?

  • jsgjsg Member, Resident Benchmarker

    Please, kindly discuss yabs results elsewhere!

  • JasonMJasonM Member
    edited October 21

    @emgh said: Only major complaint I have is that I suspect they use the same HW for regular offers, and imagine paying their regular prices and not getting better performance...

    yes, that what i wrote to their support. my client was interested in Cloud Dedicated Server. but massivegrid's dedicated server comes with vCore and not actual core (I had written a comment on their deal post here, but it seems to me that massivegrid is interested not to respond to genuine pre-sale question but only the greed seems to make quick money from the LET offer), and its same old and outdated cpus and SSD. Why pay $1000 per month for 64 vCore CPU on intel E5 with 2.10 MHz speed!! I can pay $250-$350 at Hetzner or OVH and get latest new generation Epyc/Ryzen with NVMe and global fast network. So I just dropped the plan to buy so called their cloud dedicated server. Pricing and feature wise its not competitive at all.

    and here's massivegrid's 4 year old review i found on linkedin!

    But as @jsg said, he is fine with what he got (VPS) for a 4-year period at lowest price point. Probably users will use it for vpn or some low-end work load.. considering the network, disk, server performance.

    if it improves in future, i will give it a try.

    Thanked by 1jsg
  • mwmw Member

    if this can run my unifi controller ill be happy

  • @mw said:
    if this can run my unifi controller ill be happy

    I run mine on a literal AP with a dual-core ARM A53 @ 1 GHz & 1 GB DDR3L RAM. I think you'll be right

  • ezethezeth Member, Host Rep
    edited October 21

    @jsg

    @jsg said: Now, the disk ...

    Spoiler

    Can you copy/paste the commands you used?
    I'd like to give it a little try myself on my homelab ceph setup. It's also made up of a few SSD.

    With ceph it's pretty much only the single client that will be throughput limited due to network latency. If you ran the same commands on 5 different vm for example, all should get the same numbers so long as the OSD themselves isn't at their limit or CPU is limited.

    I won't expect more than 1k iops per client, nor would anyone need any more on a 1G VPS.

    This reminds me of RAID-10 vs RAID-6 discussion. I will always prefer the latter because higher numbers = more throughput and I don't need the throughput. When designing a system you need to think of what the limiting factor is. Is it the disk?

    You'd get similar stats at AWS or Google Cloud VM. Disk is not a limiting factor.
    I dare you to run the same benchmark there :D

    Thanked by 1jsg
  • jsgjsg Member, Resident Benchmarker
    edited October 21

    @ezeth said:
    @jsg

    @jsg said: Now, the disk ...

    Spoiler

    Can you copy/paste the commands you used?

    Sure, but it won't help you without the program.

    I'd like to give it a little try myself on my homelab ceph setup. It's also made up of a few SSD.

    With ceph it's pretty much only the single client that will be throughput limited due to network latency. If you ran the same commands on 5 different vm for example, all should get the same numbers so long as the OSD themselves isn't at their limit or CPU is limited.

    I won't expect more than 1k iops per client, nor would anyone need any more on a 1G VPS.

    This reminds me of RAID-10 vs RAID-6 discussion. I will always prefer the latter because higher numbers = more throughput and I don't need the throughput. When designing a system you need to think of what the limiting factor is. Is it the disk?

    You'd get similar stats at AWS or Google Cloud VM. Disk is not a limiting factor.
    I dare you to run the same benchmark there :D

    "there" meaning the two mentioned clouds I suppose? Sure, I "dare" to benchmark any x86-64 server. And I and my benchmark don't care about Raid this or that, Ceph or others, it simply finds out what performance a user/customer on a given system can expect. Although I have quite some parameters to allow for "fine tuned" testing (e.g. "trash disk caches" or not, run normal total test size (2 GB) or pretty much any from very small to giant, etc.). But that is mostly used when a provider wants to help them to hunt down a problem, test a new config, etc.
    Arm-systems my program can't test so far but in case I get my hands on a tasty, beefy Arm server (or VPS on it) I'd quite likely make an Arm version of my benchmark too.

    You see, being a techie I of bloody course am interested in (and looked more superficially into) diverse technologies. While I wouldn't call myself an expert I think I have a pretty good idea how computers, networks, and DC work.

    But as a benchmarker I don't give a flying fuck, I only care about what a customer/user gets and can expect to get out of their VM or dedi. And just like with CPUs I'm not "religious" at all; since a couple of years I prefer AMD over intel because bang per buck is better but I have absolutely no qualms about an intel E5 v4, Gold, etc. based VPS and will happily buy and use it if it does the job. I do have some Epyc and Ryzen VPS but frankly, IMO those are overkill for most jobs.

    My opponent, if one wants to put it like that, are not AMD or intel, not this or that provider or product. My opponent is marketing BS and sales blabla and tricks. That is what I try to fight and what my benchmark is designed to detect and uncover.

  • @JasonM said:
    those 1GB VPS and other too seem oversold.
    one of my friend bought 2GB VPS and he can't even load WordPress Admin area because of 1 MB/s IO and bad network he was complaining... it seems massivegrid has old servers with old hard disk instead of SSD.
    good only for hosting private VPN and nothing else serious like production sites.

    Plus they claim they are 21 years in business and still there are not even 10 reviews or any independent bloggers posting about their vps/dedicated server usage. Google massivegrid and you'll find only information about them and no reviews since last 21 years or since 2014 (when it was started to be called Massivegrid)

    Also I found they sell Cloud Dedicated server with vCores LOL instead of pure Dedicated Cores. The meaning itself becomes useless when you see its just a virtual core (VDS) hosting and not REAL Dedicated Servers. I found that 64GB vCore so callled Cloud Dedicated will cost me $1000+ per month with massivegrid. Rather a user will go with pure bare metal dedicated server for like $150-$250 price range.

    In short, I was inclined to buy their VPS offer (buy for 3 years, get 4th year free) but it seems not useful for my use-case of hosting even a lite-weight/static production site on their so-called HA cloud vps line.

    massiveoversold

    Thanked by 1jsg
  • It is probably ok for lowendtask and/or testing...best for idling or avoiding.

  • jsgjsg Member, Resident Benchmarker
    edited October 21

    @malignify said:

    @JasonM said:
    those 1GB VPS and other too seem oversold.
    one of my friend bought 2GB VPS and he can't even load WordPress Admin area because of 1 MB/s IO and bad network he was complaining... it seems massivegrid has old servers with old hard disk instead of SSD.
    good only for hosting private VPN and nothing else serious like production sites.

    Plus they claim they are 21 years in business and still there are not even 10 reviews or any independent bloggers posting about their vps/dedicated server usage. Google massivegrid and you'll find only information about them and no reviews since last 21 years or since 2014 (when it was started to be called Massivegrid)

    Also I found they sell Cloud Dedicated server with vCores LOL instead of pure Dedicated Cores. The meaning itself becomes useless when you see its just a virtual core (VDS) hosting and not REAL Dedicated Servers. I found that 64GB vCore so callled Cloud Dedicated will cost me $1000+ per month with massivegrid. Rather a user will go with pure bare metal dedicated server for like $150-$250 price range.

    In short, I was inclined to buy their VPS offer (buy for 3 years, get 4th year free) but it seems not useful for my use-case of hosting even a lite-weight/static production site on their so-called HA cloud vps line.

    massiveoversold

    Hehe, good allusion. Seeing diverse offers and pricing from them one pretty much can't but think that they targeted the higher end business market but weren't as successful as they hoped for. So, now one can get products that usually are sold for an arm, a leg, and a kidney for a really low price, albeit with a number of vCPUs per HWT they prefer you to not know about.

    But whatever, I looked at a promo offer here and found its bang per buck ratio attractive but with a major caveat (don't try to run disk heavy jobs on it). For an NS, a VPN and many other tasks incl.a "lite" website though I think it'll do fine.

    Based of course on the assumption that that VPS really runs for 4 or 3 or at least 2 years.

  • I clicked on their thread, read it and then clicked the back button on my browser

    Thanked by 1ServerBachelor
  • and yet so many people seem to be purchasing that stinky xeon processor nearing a decade old at this point :s

  • I had one month free trial from @MassiveGRID in April this year and decided not to take this current offer.

  • Well, at least they honor my refund request (within 15 days) so I've no bad feeling towards them. Only buy if you know what you're buying.

    Thanked by 1ServerBachelor
  • Bought a VPS from @MassiveGRID prior to the LET sale, honestly pretty content with it other than the poor read/write speeds. I like that the (slow) panel has options for a firewall so I get to use something along w/ UFW. CPU steal has been alright but it's been gotten slightly worse in the past few days, blaming you LET!

    Thanked by 1jsg
  • ArkasArkas Moderator

    $7

  • NanjaNanja Member

    Hmm,

    massiveoversold?? maybe, I got that monthly plan for
    16 Cores | 48GB RAM | 512GB Storage

    Everything worked great, it's just the IOPS wasn't at what I needed.
    I requested a refund, they honor them rather fast.

    I got my money back in 2 hours 40 minutes from the time I opened a ticket.

    If you try them, you might like them. Just keep in mind you can always get your money back ^_^ :D

    Thanked by 1jsg
  • @jsg said: Pardon me? Not even 1 MB/s (write) IO and just a bit over 200 IOps? What a shit show! Hell, even the (supposed to not be particularly fast) spinning rust storage drive on my @host_c VPS does better! And this one here is sold as "SSD CEPH Storage" (quoted from OP). Yuck!

    Is this their offer that advertises Triple replicated H/A SSD CEPH Storage? If so then yes I'd expect it to be that slow. IIRC, OVH public cloud had such a setup at one time and IOPs was inline with what you're getting. You're sacrificing performance for redundancy.

  • jsgjsg Member, Resident Benchmarker
    edited October 22

    @BigDongLong said:
    and yet so many people seem to be purchasing that stinky xeon processor nearing a decade old at this point :s

    Well, maybe I'm very modest plus an old-school realist, but I still like Xeon E5 v4 a lot. But then, I'm not running a site with gazillions of requests per day/hour/second, only something roughly in the ballpark of LEB/LET.

    Frankly, I think that many, if not most, here mix up "need" and "want".

    A "stinky" Xeon E5 v4 still is easily fast enough for > 90% of all websites, with plenty RAM, halfway fast disks say, good SSDs, etc.

    Do not judge a processor based on a cheap VPS but on a dedi!

    @beermachine said:
    Bought a VPS from @MassiveGRID prior to the LET sale, honestly pretty content with it other than the poor read/write speeds. I like that the (slow) panel has options for a firewall so I get to use something along w/ UFW. CPU steal has been alright but it's been gotten slightly worse in the past few days, blaming you LET!

    Probably depending on location and distance/connectivity to it. In my case (UK, London) and connecting from within Europe I find the panel very acceptable to fast.

    @Arkas said:
    $7

    Per year? OK - per 6 months? Nuh, too expensive! *g

  • jsgjsg Member, Resident Benchmarker
    edited October 22

    @eva2000 said:

    @jsg said: Pardon me? Not even 1 MB/s (write) IO and just a bit over 200 IOps? What a shit show! Hell, even the (supposed to not be particularly fast) spinning rust storage drive on my @host_c VPS does better! And this one here is sold as "SSD CEPH Storage" (quoted from OP). Yuck!

    Is this their offer that advertises Triple replicated H/A SSD CEPH Storage? If so then yes I'd expect it to be that slow. IIRC, OVH public cloud had such a setup at one time and IOPs was inline with what you're getting. You're sacrificing performance for redundancy.

    FULL ACK. Yes, it is (advertises Triple replicated H/A SSD CEPH Storage). That's not really a smart way to go about it - and may well be a major factor in @MassiveGRID not doing so well in the higher end business market segment they seem to have targeted (disclosure: I assume but I do not know). But as the sheer endless "my order number is ..." posts suggest it seems to be quite acceptable here if the price is low enough.

    Would I even consider a more serious VPS (or even VDS) from them, obviously at a higher price?
    Nope, I would not ... and frankly not even a super-cheap storage server. For that, the "get both, a very low price and really decent performance" provider already exists -> @host_c.

Sign In or Register to comment.