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.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
I see what you suggested there. Right.
That'll shut it right up.
It was the first time that I had tried NetBSD without a swap partition, and by default NetBSD really wanted to see a partition /dev/wd0b, even with no corresponding line in /etc/fstab. Kind of surprised me at first, but then I found a workaround (not yours!).
Can NetBSD even address 10GB of RAM?
Since December 9, 2004.
(Okay, Linux was quicker in this respect.)
From what I've been reading , It's not a bad idea to have such a tiny swap for compatibility reasons. I use 256.
This was also my initial inclination (but with 128 MB), but I'm not sure that compatibility reasons are still a valid justification (though this may depend on the system and what is meant by "compatibility").
My reasoning is that with 10 GB RAM and 40 GB disk space, disk space is a relatively scarce resource compared to RAM, so why sacrifice even a little of it to swap space if the swap is unlikely to be used? (At least in my case for this server, given that I'm not running a lot of heavy applications all at once.)
128MB with vm.swappiness = 1 sounds about right, in my book.
Create a swapfile, if not a partition and allow *nix to do its' memory management.
The lack of disc space prevented me buying this one - phew! One less idler.
Maybe I'll throw NetBSD on that VirMach 10GB/10GB box. I can't think of anything else to do with it other than use it as the worlds' possibly slowest external memcache. I know they've got OpenBSD and FreeBSD ISOs, not sure about NetBSD.
could someone tell me what the results are of "sysbench --test=cpu run" on this?
Does any one know how much it costs for extra Storage?
@MikeIn
@hjlow
Yes, this was what I used to believe as well. But swappiness =0 no longer behaves as the manual says in the last few Ubuntu LTS kernels. YMMV.
So, I'd use a 128M swap and swappiness =1 (which oddly somehow behaves better than 0)
There were quite a few writeups about this phenomenon in the tech blogosphere.
Ubuntu breaks a lot of shit. Mostly Debian.
I don't recall if this was a change in mainline or canonical TBF
If I recall well, the new installer introduced in Ubuntu Server 18.04 LTS (and used in the standard installation ISO) doesn't offer one the opportunity to configure a swap partition (but it can use an existing swap partition). (!)
My personal recommendation is to install Ubuntu Server 18.04 LTS using the special installation ISO that uses the installer from Debian, which allows for all of the standard options.
Oh, you managed to get one of those unusual VirMach boxes.
If you need any help with NetBSD, let me know, but I think that you know what you're doing. Any of the BSDs would also be a good choice for that VPS. You could be brave and try MidnightBSD. (On the other hand, TrueOS would probably be overkill.)
if only it were hdd
Yeah that's the issue, and it becomes even worse when you have several idle machines as they cost the same as a good quality VM with a top noch provider: you end up hosting your website on one of your oversold VMs while you could have used something better.
Of course, that's not 100% true, some bargains are great and some providers offer great quality, even for their specials, but when you start having more than an idle VPS, cheap isn't necessarily so cheap anymore, and too good to pass often not much more than an illusion if you look at the bigger picture.
Drinking, and Virmach specials seem to go hand in hand. Bastards.
If there was decent HDD, I'd probably throw Dragonfly on it, because HAMMER2 intrigues me. Plus, I knew the dude from his Arkanoid clone. Kind of.
How's the CPU on that one? I know that you can't use more than 50% for x minutes or the VPS gets shutdown, but if it has enough cpu power, it could be a cost-effective VM to process files and store those in a redis db (api backend for example). Good luck!
You make some good points. With the VPS in this thread, I intend to replace a couple of small VPSes.
Well, I re-added the swap partition (128 MB). It's probably good to have one (especially if I want to use 10.01 GB RAM), and it makes NetBSD happy.
Missed this one again, I will blame my boss more. -_-
I can transfer you mine if @PHP_Friends allows it. You'll have to pay me though
P.S. Netflix works on these. At least my friend's one did
Something isn't right here. E5-2630 has 12 threads, taking at least 1 for hypervisor/host leaves 5 guests. 6euros x 5 = 30euros a month revenue per physical server?
I think that v4 has 20 threads (or am I mistaken?), so using your reasoning, this would make €6 x 9 = €54/m.
It's also possible that the promotional VPSes are distributed across machines, in which case there are also normally priced VPSes on any machine.
But you ask a good question.
Probabilistic overcommit on CPU is possible given enough host nodes to spread guests around on
As this was a limited offer there are dozens possibilities. For example as marketing promotion or to cut losses with unused resources.