Howdy, Stranger!

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


Netcup's Advent calendar / Adventskalender - Page 2
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.

Netcup's Advent calendar / Adventskalender

2»

Comments

  • Cloud vLAN Giga ADV20 - 5,99€

    https://www.netcup.eu/bestellen/produkt.php?produkt=2694
    https://www.netcup.de/bestellen/produkt.php?produkt=2694

    I don't really see why they make one pay for this but well, I guess it might be interested for some.

  • I'm really quite confused about Netcup's reputation. I've heard that they're both wildly oversold and overprovisioned with relatively average performance, and others seem to hail Netcup as this great bang-for-the-buck provider.

    Maybe I'm not reading broadly enough but is Netcup considered to be premium, average or an avoid?

  • @Bertie said: average performance

    Netcup's price-resource ratio is very good and they are very dependable in terms of uptime/availability/network.

    If there's anything that is a bit poor (or lowend) as far as their offerings go, it is usually their VPS lineup where the CPU is definitely not the fastest on the block and the performance varies quite a bit. As the nodes get loaded/balanced, the CPU oriented benchmarks will show degradation.

    Their RS (Root Server) line up is much better (they offer dedicated vCPUs/Threads) and the performance is much more consistent but again, it is not as good as it can be for the given hardware specs (and this is likely due to throttling/balancing/limits in place to protect the rest of the (ab)users). Also note that they don't have raw/full CPU pass through (and limit many flags not just for virtualization).

    Their IO (esp. on the SSDs) and network is really very good and consistent.

    Note that depending on the kind of work load, you wouldn't even notice anything amiss as far as their VPSs go because it is more than sufficient for your typical web server type of needs with generous memory to go with it.

    Beware of their contracts and cancellation policies - read the fine print thoroughly.

    Thanked by 3Bertie Falzo pbx
  • @Bertie said: Maybe I'm not reading broadly enough but is Netcup considered to be premium, average or an avoid?

    They are premium of course, but as their customer u need to read their General Terms & Conditions of Business (T&Cs).

    I have received printed-mail invoice into my house (ID) when i didn't pay where the contract still go on. You can search keyword "debt collector" to know.

    Thanked by 1Bertie
  • bought 4 netcup cups

    Thanked by 1Falzo
  • FalzoFalzo Member
    edited December 2020

    @hzr said:
    bought 4 netcup cups

    .

    @nullnothere said: they are very dependable

  • @hzr said:
    bought 4 netcup cups

    already sold out, unfortunately :(

  • @hzr said: bought 4 netcup cups

    I fully believe in failover and load balancing for the morning cuppa. I mean, if you don't get the cuppa dependably you can imagine how the day goes south quite quickly.

    I'm not so sure on the crash-proof-ness of a ceramic/porcelain cup (even from Netcup) though.

    @Falzo - if they had an unbreakable cup...

    Thanked by 3Falzo pbx sportingdan
  • @Bertie said: is Netcup considered to be premium, average or an avoid?

    I agree with the answer @nullnothere gave you.

    I'd say they are average as far as 'pure performance for the buck' goes, but premium in network and seriousness: if you want an host you can trust to stay around for a long time, to have good (security, technical, etc.) practices and to offer you stable hosting on a good network, it's a great choice.

    Sure you can find better/cheaper bench monsters from lower end providers, but sometimes you don't want the best bang for the buck on that front but a serious and very stable provider. Netcup is pretty well suited for that IMO. If you know how to properly use your RAM (mysqld caches, varnish, redis...) their HDD VPS are really nice machines, for a damn good price...

    Thanked by 2nullnothere Bertie
  • BertieBertie Member
    edited December 2020

    RS Mistletoe ADV20
    Der Mistelzweig darf in der Weihnachtszeit nicht fehlen. Mit unserem RS Mistletoe haben wir etwas ganz Spezielles hinter das 9. Türchen gepackt!

    AMD EPYC™ 7702 (8 dedicated)
    48 GB DDR4 ECC RAM
    KVM-Technologie
    99,9 % Mindestverfügbarkeit
    Sicheres und schnelles RAID10 (Hardware-RAID-Controller)
    Konsole zur Fernwartung, Backupsystem uvm...
    

    30,30€ pro Monat

    Thanked by 1pbx
  • hacktekhacktek Member
    edited December 2020

    Netcup is good bang for the buck, I've used them for a couple of months but I have run into a very annoying issue that support has not been able to fix. Whenever my box goes into heavier load it tends to be relatively io intensive. What I've noticed is that, during heavier loads, the IO performance will start to gradually degrade until the box becomes unusable. An OS reboot does not help, only a control panel power cycle. Here's an example of what I'm talking about:

    https://imgur.com/IR783jm

    You can see that IO started to spike due to the heavier use and then you start to see the it ladder down until it literally flatlines. Here'a a closer look of it flatlining until I power cycled:

    https://imgur.com/nXeUEIr

    I'm about to be done with this heavier type of work and the box will resume its regularly scheduled jobs which aren't as intensive (at least they're more spread out) so I don't mind keeping this server because of the price but support is clueless...I'd say that this is some sort of rate limiting but it doesn't sound like it is based on the fact that the problem goes away after power cycling...

  • @hacktek said: IO performance will start to gradually degrade until the box becomes unusable. An OS reboot does not help, only a control panel power cycle.

    Interesting. Is this on a "special" plan or on a VPS on their regular line? Did you try waiting to see after how long it came back to normal? It would be good to know if you have to power cycle the VPS, even after waiting for some time...

  • xetsysxetsys Member
    edited December 2020

    @hacktek said:
    What I've noticed is that, during heavier loads, the IO performance will start to gradually degrade until the box becomes unusable. An OS reboot does not help, only a control panel power cycle.

    Exact same experience here. Perhaps there is a reason why the contracts are in place so that customers don't leave them after finding out their service is not cut for sustained load. Good for benchmarks though. I have 02 root servers, one with HDD & SSD & a VPS with SSD.

  • @xetsys said:
    Exact same experience here. Perhaps there is a reason why the contracts are in place so that customers don't leave them after finding out their service is not cut for sustained load. Good for benchmarks though. I have 02 root servers, one with HDD & SSD & a VPS with SSD.

    You can get a refund using their satisfaction guaranteed thing. I got out of a 12-month contract with that.

    The messed up the invoices though so u gotta contact support. If you paid via PayPal also need to contact support because the default option is to receive it through bank.

  • @xetsys said: Perhaps there is a reason why the contracts are in place so that customers don't leave them after finding out their service is not cut for sustained load.

    I don't think that they are looking to scam/deceive their users, contracts aren't there to keep you with them if you are not happy IMO.

    You had this problem with the VSD line? What kind of load are we talking about here? Maxing out i/o for hours?

  • xetsysxetsys Member
    edited December 2020

    @pbx said:

    @xetsys said: Perhaps there is a reason why the contracts are in place so that customers don't leave them after finding out their service is not cut for sustained load.

    I don't think that they are looking to scam/deceive their users, contracts aren't there to keep you with them if you are not happy IMO.

    You had this problem with the VSD line? What kind of load are we talking about here? Maxing out i/o for hours?

    downloading isos and uploading to google cloud without encryption. Talking about rootserver w/ SAS storage here. The HDD is actually pretty fast but after like an hour, download cripples due to high IOWAIT. During that moment you realize the traffic limits are just a gimmick since one can not even exhaust half or quarter of that. Or, may be if the server is loaded with large linux ISOs with multiple people leeching 24x7. I have tried changing kernel and distribution, but it seems like that there is some kind of QoS feature that does it. Internet search shows that the disks are qcow2 type instead of RAW so, the performance degrades once you start loading them. Restart from control panel SCP does fix it but the cycle continues. The restart from terminal strangely doesn't have any effect

  • @xetsys said:

    @pbx said:

    @xetsys said: Perhaps there is a reason why the contracts are in place so that customers don't leave them after finding out their service is not cut for sustained load.

    I don't think that they are looking to scam/deceive their users, contracts aren't there to keep you with them if you are not happy IMO.

    You had this problem with the VSD line? What kind of load are we talking about here? Maxing out i/o for hours?

    downloading isos and uploading to google cloud without encryption. Talking about rootserver w/ SAS storage here. The HDD is actually pretty fast but after like an hour, download cripples due to high IOWAIT. During that moment you realize the traffic limits are just a gimmick since one can not even exhaust half or quarter of that. Or, may be if the server is loaded with large linux ISOs with multiple people leeching 24x7. I have tried changing kernel and distribution, but it seems like that there is some kind of QoS feature that does it. Internet search shows that the disks are qcow2 type instead of RAW so, the performance degrades once you start loading them. Restart from control panel SCP does fix it but the cycle continues. The restart from terminal strangely doesn't have any effect

    It really sounds like you're being rate-limited, where the rate limiting is tied to the current session from the perspective of the control panel.

    It's hard for us to determine -- based on what you've said -- whether the rate limiting in your case is justified or not.

  • it's a shared service anyway. at least disk and network is...

    if you think you can be the only one torrenting the sh*t out of it and disrespecting the fact that your neighbours might hope for a usable portion of the ressources for themselves, I can only say I am happy they have some QoS in place to balance that properly.

    better move on to another provider, large linux isos blabla with a german provider is usually not a good idea anyway. they will quickly show you the door if you manage to get DMCA notices or cease&desist letters sent their way.

    Thanked by 1angstrom
  • xetsysxetsys Member
    edited December 2020

    @Falzo said:
    snip

    There is no way Netcup will get DMCA. These are private ISOs. Netcup should advertise properly what they are selling. That's the crux behind my complain. 80TB or 120TB is a misleading figure when the user will be capped to 100-200Mbps unless they get around to it by rebooting server every hour. It's a root server and not VPS. It should be able to meet the traffic quota before rate limit.

  • @xetsys said: There is no way Netcup will get DMCA. These are private ISOs.

    'private' usually does not make seeding legal ;-)

    anything in conflict with the law makes you breach their ToS, no matter if they get letters or not. so if you don't get what's advertised while breaching their ToS, at least you shouldn't complain... but that's just my humble opinion. :D
    these boxes are clearly not advertised nor intended to be used as seedboxes.

    more likely you can use the traffic or network speed advertised if you don't use software that eats up IO that much and makes you run into limits to balance the hostnode.
    of yourse you won't be able to reach higher network speeds than what you can read/write to your storage.

    as said, I guess they are simply the wrong provider for your use case 🤷‍♂️

  • @Falzo said:
    I guess they are simply the wrong provider for your use case 🤷‍♂️

    Hetzner it is, thats what I figured.

  • @xetsys said:

    @Falzo said:
    I guess they are simply the wrong provider for your use case 🤷‍♂️

    Hetzner it is, thats what I figured.

    best of luck!

Sign In or Register to comment.