All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Contabo: a warning
I think I am done now. I purchased the "double SSD" version, 8 GB, four cores and 400 GB SSD space. The experience has been painful.
I ordered Rocky Linux. After extreme sluggishness, timeouts and various frustrations, I reinstalled it. After the reinstall it was fine for an hour. Then it started to get slower and slower again, until it was unusable. I disabled all imaginable services, leaving practically just sshd running, and rebooted it twice. No change. The disk speed dropped from 330 MB/s to 20 MB/s.
Running simple commands like
certbot --version
Took from two seconds up to one minute. Launching anything felt like having your file system on a floppy disk.
Wanting to test whether they have a bad image for Rocky, I installed Debian to see if there is any difference. Well, there wasn't. Installing Sendmail, for instance (just to test) took almost three minutes.
Start-Date: 2022-02-16 12:15:28
Commandline: apt-get install sendmail
End-Date: 2022-02-16 12:18:15
Disk speed was marginally better than in Rocky
dd if=/dev/zero of=/tmp/test1.img bs=1G count=1 oflag=dsync
1+0 records in
1+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 28.8645 s, 37.2 MB/s
You can get a glimpse of the joy from this test:
time certbot --version
certbot 0.31.0
real 0m19.342s
user 0m0.997s
sys 0m6.389s
On an immediate second launch, it is just 0.5 secs, running from cache. The SSD speed is truly awful.
As for web responsiveness, it would take several seconds to get a "hello world" to the browser, with occasional timeouts.
When I reported the problems to Contabo support, they told me they did not understand what I wrote. That may well be the case.
I have no clue about the root of the issues. Maybe they have just oversold the nodes, maybe the SSDs are far from professional quality, who knows. The server was supposed to be a secondary mirroring backup for an application server, but I have to give up now.
Comments
Completely fair and honest question: do you find the way to the toilet your-self or do you need a guide?
Contabo is insanely good for the given price. Performance, uptime. Support and network is to be desired, but I pay peanuts - I accept monkeys.
TBH 20 seconds to run certbot —version is a bit too slow in my opinion.
Of course the OP shouldn’t have high expectations for the price, but I do think the provider should also provide a usable service as well
Wait wut?
Try asking to be moved to another server. With contabo, the cpu and disks are expected to be slightly slow but you get way more than you pay for (most of the time). Are you getting these same results consistently all the time?
Maybe you are writing too much. Tell them running a simple command takes a full minute with a new installation.
Just a clarification: I am currently running 5+ KVM based servers, and I have installed "a few" Linux versions since 1994. I reinstalled the Rocky instance because there was nothing else I could try out any more. I even removed services that are running perfercly in all other instances. The last straw was to remove incrond and lfd which I can live without but even that did not help. Debian was snail-slow right after the installation, with no self-installed services running.
If the wise guy above (LTniger) has any real advice to give, he can rise from the toilet seat he seems to love and give that advice - or wait until he is fourteen years old.
which location?
It should be Germany (38.242.207.20). The location was not critical to me, however.
FYI: I installed Debian simply to see whether the Rocky image they had could be the culprit. Nope. The problem is somewhere deeper.
The real advice (I just shat and not anymore on the toilet) would be to change providers immediatelly and end your suffering by requesting refund. Sounds logical?
You probably shouldn't mention your IP here. Anyway, has the server been slow for the last hour? Or for the last few days? In any case, if you're really unhappy just ask to be moved to another node and see if it helps. If it doesn't help you can cancel it.
delete your ip first
i have another location and use rocky. so far fine. you should contact them and ask to move to another server
looks like it doesn't just happen to you. check
https://talk.lowendspirit.com/discussion/3770/contabo-400gb-free-ssd-if-you-order-today/p1
Contabo really depends on if your VM's root server is "noisy". If they're not, your server will be fine and will be in a good state (at an unbeatable price).
However, if they're one of those "crowded" nodes, your VM might suffer a low IOPS or whatever (It's funny how some SSD VMs are slower than a shared HDD VM that I own). If you can, you might try asking for support to move your VM.
And of course, you get what you pay for. You wouldn't expect things "blazing fast" using services that are very budget-friendly like Contabo.
@LTniger said:
i have always known you to
good luck with contabo!
Where is YABS?
Looking fine for me
That does not look fine to me.
Strange. I've benchmarked quite a few Contabo servers and also have multiple (paid for) private VPS with them, both NVMe and SSD.
What I can confirm is that Contabo seems to have limited IOPS on their SSD VPS (to a a reasonable level) and that they seem to have a problem wrt node/VPS monitoring that results in somewhat of a lottery whether one happens to end up on a good (no or few abusers) node or one plagued by abusers.
What I've never experienced is what OP describes, both in terms of systems being ridiculously slow and in terms of technical incompetence and ignorance of their support. As long as @mkv doesn't show us the communication with Contabo support I'm having a hard time believing what he says.
Based on my experience I'm confident that Contabo support at the very least would have a quick look at OP's node to check whether there's some problem.
This one right here, I have a few servers with them and I was probably lucky enough to get some less crowded nodes for all of them, the uptime has been at 100% for the past 2-3 years with no network issues nor disk speeds (currently anywhere from 400 MB to 1GB/s) after I've requested for the hard cap to be removed ever since the servers were delivered to me. I also only needed their support once when they had some network issues within the first days I've joined them and in my case it was extremely fast, didn't need them ever since though.
I'd also suggest asking them if they can move you to a different node, although it sounds the issue may be different. What did support tell you so far in regard to your problem, assuming you wrote them back after their initial response?
At this point you just have to accept that there are occasionally cases where contabo VPSs are ridiculous slow due to abusers on the node. I heard of contabo since a few years back only because of the reputation of contabo being cheap and allowing people to slam the cpus 24/7. And there are many people on contabo who do slam the cpus 24/7. After many years, most likely contabo limits users somewhat but these abusers still exist.
However, I am sure you do get more than what you pay for with contabo (most of the time). You need to stop implying that complaints about contabo are all mudslinging attempts. All op needs to do is probably request to change to a different node. The occam's razor explanation about the support's reply is that op simply wrote too much useless information in the ticket and support couldn't be bothered to read his wall of text.
I'm not implying or assuming that complaints about Contabo are all mudslinging attempts. In fact, I myself used the term "node lottery" and I know that Contabo doesn't reign in sufficiently on abuse, probably (my relatively educated guess) because their node & VM monitoring and management doesn't allow it. ("Funny" side note: And if they actually did start to reign in on abuse LET quite likely would be flooded by "Contabo brutally limiting performance!!!" threads ...).
However, experience (not only) on LET teaches that when someone opens a thread in order to complain about this or that and basically just bashes a provider, chances are that it's about mudslinging.
Frankly, looking at OP's problem description and style my guess is that Contabo support did read his email, did look at the node, didn't find a problem, and couldn't make sense of his weird description (as in "certbot and whatnot allegedly slow? Me not care, the node is OK and no significant abuse noticable. Done").
That said, sure, it's possible that OP ended up on a really, really, bogged down node but before condemning Contabo I'd like to know a bit more and to see some evidence.
Plus, why didn't he simply ask to get on another node? Opening a mud-slinging critical thread on LET certainly didn't need less effort than writing a single line email to Contabo's support ...
I did have the exact same issues with response sometimes up to a minute or two for some commands.
I presented the fact and they didnt understand what I meant even though I showed them multiple time things like you have.
Then I got mad and presented about 4-5 tests and and showed them again how bad it was, sadly I didnt check the CPU steal back then so check that as well.
It took about 4 days and they moved me to another hypervisor and it worked out ok for some time but I started to see same shit about to happen again.
Not to mention that they moved me from Ryzen to an old Xeon chip, but to be fair they where both shit in performance
Both of the VPSes I did have there are canceled and Im very happy with that and never looking back.
Got much much better performance at the VPS I got when Inceptionhosting had a 24H sale, even with less then half the CPU cores and RAM.
Ask them to move, it may take a few days before it happens and then its just a waiting game to see when that shit starts preforming bad and move again
Or just accept that those few Euro are somewhat wasted and move on to something else, everything else I have tried have preformed better so far
GLHF
If you want a fast Contabo VPS, get an NVMe one rather than an SSD one.
I had up to 30% CPU steal on my NVMe, but sure the disk where indeed faster.
Maybe my one was an outlier? I had a NVMe VPS in Seattle for a few months and CPU steal was usually under 5% (albeit with some spikes to 6-8%), and it felt very fast the whole time I had it, even when CPU steal was elevated.
Well can't talk about US servers, my experience is from the German ones.
And I used to have i both DC in Germany.
As it is a German company I would guess most customers are within EU, so let's just hope US one and wherever they may have not go down the same road as both German ones.
They provide a high storage space size so in order to keep their servers away from backup and storage abuses they limit the disk speed, they does for the CPU speed too for sure but I’m sure about the disk, one more thing I think they managed to trick the benchmark scripts to show higher results because even the results are way more than the actual speed.
If you're mentioning about my post, I know the 4k is quite poor, but it's fine for my use case, at least it's better than OP's.
I remember I saw on somewhere that contabo"s Germany server had quite poor performance, But the US is fine, maybe you can try again next time. @mkv
Always more stable
Well, from my personal experience, I have never seen such slow VPS as at Contabo. I gave them several chances, but every time I was disappointed.
The worst my experience was... VPS S SSD (8GB RAM) as a mailserver made with help of iRedMail.
Normal memory usage was ca. 5GB and I had some crons to renew certificates, as I was hosting several domains on one mailserver (including different MX).
And one night, I received an alert from my monitoring, that there are issues with certificates.
What exactly happen? All crons have been set up to work once a month, between 01:00 and 1:03 UTC+1 and they failed. For unknown reason, server hanged up and rebooted, or something disturbing happened. One of the cron just DESTROYED whole mailserver, update failed, certificates expired etc. And the worst thing was that due to mess that has been done by this reboot, I was unable to make this server fully operational again. Luckily I always take backups.
I immediately contacted support and... they didn't know what happened - I tested scripts and their memory usage and I am 100% sure, that I did not run out of resources. Support said they optimized my server (LOL?!), but next month the same issue occured.
According to iRedMail docs, 4GB is enough for low traffic production server - I had about 6 accounts registered there - this IS THE LOW TRAFFIC SERVER.
This is my experience here... I am not power user when it comes to CPU, so for me, CPU can be shared between 69 customers, I don't care. when it comes to RAM, mailservers require RAM. And I cannot accept overselling resources the way that causes random shutdowns, when I am not using at most 6GB of RAM on 8GB RAM VPS!
Also, I have never experienced such slow SSD on VPS. My RAID1 made of two HDD 5400RPMs is faster...
Summarizing, I gave Contabo many chances, because maybe I was being placed on overloaded nodes. But every time, after short time, the same issues started to occur...
I moved Mailserver to Netcup - now I can say it's the stablest server I have ever experienced.