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.
kvm 2xE5-2680v4/2gb ddr4/50gb nvme+3.6tb hdd raid10- located in Bucharest ! Telia upstream !
This discussion has been closed.
Comments
Same , PMed . Hope you can spare some time to process, thanks~
Nope, I'm a big lair.
I looked into it:
18/03/2021 PROMO 18.03.2021 $37.50USD
18/03/2021 Add Funds Invoice # $75.00USD
Thanks guys!
My original NVME speeds have increased from 2MB/sec download to 12MB+
B
Hi coicu, I cancelled my vps a couple days ago, it just shows "terminated" but not yet "cancelled" So if you think you really solved the HDD issue, I am willing to abort my cancellation request and give it another try if you can explain what the HDD problem was and how you solved it.
Thanks in advance.
I have another storage server from cociu with much slower cpu (only 1 core) and 1gb of ram and its way faster than this one...
hmmm, interesting... Which means this one really had issues and not just oversold/ overloaded by other users. That's what I tried to point out a few days ago, The performance was lower than just some users abusing it imho.
p.s: before termination of my VPS I noticed that I when I download a file for example with curl or wget, I get ~100MB/s write speed on the drive for the first 20-30 seconds, then speeds start to go down until reaching 1-2 MB/s and stay slow even when I start new downloads for a while after it
Ticket: 29/03/2021
PM: 07/04/2021
Any reply would be much appreciated
Thought it might be USD/EUR conversion but seems it worked quite well in USD for _FBi
You're reading mine backwards, but yeah, they nailed it. I just missed it.
I do have a sales ticket collecting some dust. Nit a bad thing, as it prevents me from spending more haha.
Dear cociu, my service was still in pending state. I paid the service since 26/02/2021. Can you help me to check my ticket please? I really want to use this vps service. Thank you very much.
Hm. Since today my VM freezes completely anytime there's some I/O on the storage.
Anyone with the same kind of problem?
Same here too, any time I try to download something to the 3.6TB HDD. Once in a while, an error message gets printed to the display (visible through noVnc), but I keep forgetting to screenshot it.
`fio Disk Speed Tests (Mixed R/W 50/50):
Block Size | 4k (IOPS) | 64k (IOPS)
------ | --- ---- | ---- ----
Read | 4.79 MB/s (1.1k) | 2.33 MB/s (36)
Write | 4.81 MB/s (1.2k) | 2.50 MB/s (39)
Total | 9.61 MB/s (2.4k) | 4.83 MB/s (75)
| |
Block Size | 512k (IOPS) | 1m (IOPS)
------ | --- ---- | ---- ----
Read | 4.33 MB/s (8) | 9.27 MB/s (9)
Write | 4.84 MB/s (9) | 9.89 MB/s (9)
Total | 9.18 MB/s (17) | 19.17 MB/s (18)
`
noice
Is this the speed of NVMe?
@cociu
Can you check your inbox? Thanks!
Something like this?
Did you do anything about it, yet?
For now only a cold restart "helps" for some seconds up to minutes.
I've had the same problem of the VPS freezing for about 4 days. It only happened when writing data (ftp) to the 2nd disk (didn't test the 50GB disk). Before then I had been able to ftp data to the 2nd disk without problems. Reading data has worked the whole time. I don't know if this is a coincidence but I changed the IO scheduler to none for sda and vda and have now been able to ftp more data without any freezes.
That really seems to work. Thank you!
Weird that this is working, instead of mq-deadline.
( Looking at https://wiki.ubuntu.com/Kernel/Reference/IOSchedulers )
Have a try @jabashque!
Setting the scheduler to "none" for a spinning hard disk somehow feels wrong to me, even though that actually should be what we should be setting for all our virtual disks since running any I/O scheduler in the guest is ultimately just a waste of CPU cycles.
In any case, thanks for suggesting this! I haven't run into any issues since after setting the I/O scheduler for the 3.6TB to none.
EDIT: seems like it works fine. My only concern is that I don't understand why the VM hangs if I try writing something to the 3.6TB disk when using an I/O scheduler. I don't know if that's a sign of a bigger underlying problem.
If you are referring to especially those "interrupt took too" messages, there is nothing to worry or do about.
IMHO the verbosity level of these messages is a bit inaccurate as they are mostly only informational and the kernel is already doing everything needed to properly handle them.
(cf. https://askubuntu.com/questions/621490/what-does-perf-interupt-mean)
Thank you for sharing - This setting made my seedboxes much more responsive
Well, changing the scheduler to none seemed to work but a few days later I got a lock-up when trying to ftp some data - after less than a minute. As a test I've just copied 9GB from sda1 to vda1 and it's taken 6 minutes. Very frustrating. No point opening a ticket because their support is non-existent. Refund or chargeback is looking certain.
i think the best way is open a tiket and ask refund ... but finnaly is your choise
Just requested a refund in ticket #504862 @cociu
Not sure whether others are seeing better speed, but hdd didn't improve beyond ~4MB/s for me (read/write), even with scheduler fix
@cociu Can you check your inbox?
Maybe because server and storage are in two different location
None of my tickets/messages got answered for weeks...
I got my refund after messaging through PayPal by opening a case.
does this mean there will be some open spots? i want one
Yeah, it seems to be getting worse; the chance of your VM locking up and freezing when writing to the 3.6TB disk keeps growing bigger and bigger, regardless whether you've set the I/O scheduler to none. I'm more likely to guess that it's a hardware-related issue on the host machine at this rate.
I swear to god, if that's the real reason...
No, not you, others are plotting. The resources are just shared. It's what we call "noisy neighbors".