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
Hello, first thanks for reply.
apologies but u r not right about virtualmin, it
s not truth what u wrote, and i cannot agree: i have rich and longterm experiences w virtualmin (my strongly preferred server GUI CP) : recommended/required minimum for --minimal version is 256MB, necessary minimum for stable running. (I have installed several instances w 256 MB RAM & 256 MB or 512 SWAP and can sure u - if node is working well and its not oversold, then this work without issues, only a few MB are free in ram and swap but its can works and stable, in accordance with condition if running websites are not hard difficult for system resources and consuming very much ram (for example mean 2-3 SEO microsites or WP as is my case, not massive/robust dynamic database webapps w hundreds db queries in each second))Here below is screen of last one at other provider installed by me also in day when i bought account at you (- 2 days left). There is only 256 ram & swap, but virtualmin is loading w/o any issue or waiting, and running quite fast and work stable. Yet On your VPS i have as resources 256 ram & 1000 swap, but its does not able load main page in browser, due from degraded server and network performance. second screen is bench: left is my vps on your cloud node before approx 8 hours back, right is freshly made bench from mentioned 256 ram vpsgod annualy $8 oVZ (their cheapest VPS, not cloud w SAN or DAS - only VM machine w local disk directly on host node) - u can compare them.
if u will have done checking and node working back normal then i can check and confirm virtualmin will able load normally and work.
regards miu
BTW: if images are very small for view in this page, then pls open them as "image on new tab" or as below mentioned URL in new tab, for fullscreen width and view in page, thanks
http://optima.space/00-public-web-imgs-890428095/virtualmin256.jpg
http://optima.space/00-public-web-imgs-890428095/virtualmin256-2.jpg
I/O Speed( 1.0GB ) : 34.3 MB/s
I/O Speed( 1.0GB ) : 29.7 MB/s
I/O Speed( 1.0GB ) : 33.0 MB/s
Average I/O Speed : 32.3 MB/s
IOPS ...
try again, ohh...
May be the allocated nodes are bad.. I have a vps with them am getting decent I/0 speeds
more than 2 days VPS is still terrible working without change (morning i tried whole reinstall on new, yum update + virtualmin install. took more than hour! and after install is Virtm. still not able load in browser dashboard page - eternal loading without end..), but later evening this happened see images below..
now i have immediately opened ticket and now waiting for their explanation and reply..
BTW: i
m very surprised from "we had to take manual intervention" so this is reason why i post it here, because when u bought VPS, paid, its not working well (2+ days) and then provider treat you w / try terminate your service (and account) from fictional reason ("you are abusing CPU resources" - when you are not using this disaster ever...) then it`s SCAM w purpose toss you over the deck without refund..i have a quite large amount of VPSes but i was never abusing any resources and never had similar issues w any other provider ever...
And now i must admint I'm full of teeth of this stupid game.. so i try last time aks them for fix, if not then i require full refund
and when i tried load details of VPS (manage buton) and check/see admin panel graphs of utilization
and this is what i see second day
as u can see: yet i do not have created and running here any website and server is empty (from simple reason - it's not normal working, not possible add and run anything here ever), but they wrote me "i did overload their CPUs long time period"...
with Virtualmin URL triggering attemp in browser??? or when SSH client was every time try always reconnect when connections was today hundreds time around losted??
That's bad for Xmas.
Luckily I couldn't wait for deals and took the ultravps yearly offer from LA.
We tried to reproduce your issue on the same config as you were, 256 MB RAM, 10GB disk and here were the results. To make it fair, we have deployed this server on the same node which your current server is on.
Server OS: Ubuntu 16.04
Time taken to install Virtualmin (256 MB RAM VM): 40 minutes (SWAP was almost full when install ended)
The page loads acceptably fine on the 256 MB VM, here is a link for testing: https://173.82.187.176:10000/ (do Ctrl + Click) also http://173.82.187.176 (do Ctrl + Click)
Also, you were referring to SSH timeouts, to test that, i have added a monitor on https://nixstats.com/report/5c1f2f6f3f0bc1389f34ce85 to review uptime for this 256 MB demo VM.
Rather than a LET exaggeration post, things would be much easier if you raised a ticket with us simply mentioning "Hey, can you migrate me to a different node, i am experiencing few issues on my 256 MB server which cannot run AWStats, Apache, Bind, ClamAV, Dovecot, MySQL, ProFTPd, Procmail, SpamAssassin, Usermin, Webalizer, Postfix, Fail2Ban"
If you require a refund, we would be more than happy to process it for you. And NO, we do not scam clients.
Regarding disk, we advertise high availability and automatic failover on the newly provisioned VMs. Which means, when you write to the disk, it writes to 2 copies at once, which costs IO (one local disk, and one disk over a 10Gbit network dedicated for storage)
After several hours, as I mentioned, the RAM of the server is exhausted with time, causing insane amounts of SWAP writes which result in disk IOPS (Mainly writes), in terms slows down your server. I would suggest to opt-in for more RAM if you need to run a stable machine with Virtualmin as it has over 20 services running which need more memory in the long run
Note SWAP is 99% full ^^
P.S The above test servers were taken down.
I recommend an additional swap partition for the swap partition.
yes, add 4GB of swap and load more services on the spinning rust.
Just use swap as your last option. Don't use it as additional RAM to avoid headache
could you solve the IO speed problem(on node 3, 8, 12), sir?
Can you raise a ticket? We don't have any nodes with IDs 3, 8 or 12
update: based on my ticket req, server has been unsuspended
usually what i`m using only in Virtualmin is httpd + MySQL + apache + firewalld + ssh = only websites hosting, time at time also bind, rarely dovecot + postfix and very rarely spamassassin, absolutely never clamav (biggest resources consumer from this package)..
now when server is online i will try disable all except apache php myslq (this is what i need have live in this case - on this vps), and then will see if it will able work. If not then i`ll ask for kindly refund via ticket
my last question - r u able to explain me how it possible that virtualmin at many (most other) providers (w disable spamassasin and clamav) is able long time and stable works without one issue from config 256 ram & swap only, without swap or other resource fatal halt (for example above mentioned vpsgod 256 ram swap ovz - still is it working and fresh running, w the same utilized ram and swap as is on picture from install day), and here not?? this is what i do not understand
and sure it`s last what i want - slander or damage any provider w/o relevant reason , but i was so many scammed and/or robbed about a lot money by "providers " advertising on LET and LEB with total shit services (as the winners in this category i can mention woot, alpha, hostmybytes, nfp, sparkvps and mega scammers deadpool dedicenter, etc.) so when i have feel things go again bad then i post issue here due from reason let know other users be aware in advance before their decision based on "appealing looking" offers and avoid also useless waste their money or/and time when happened it me
simultaneously i always glad leave kindly words on stable and honestly provider at every occasion on LET, glad support them and really highly appreciate their hard work, each time. And write rather positive suggestions and experiences as write about issues and headaches
regards
we agreed together on refund, based on my request
i would to glad confirm important thing:
they did accept my refund request easy, immediately and without useless delaying and resistance / hassle free, what is appreciated from my side and serious sale politic from their side
so now i expect refund (money receiving) within 7 days as they wrote, and end comments and updates in this thread, thanks for all interested for their suggestions..
Great. Next time dont try to run all that shit on a 256MB VPS.
thank you Mr. Wisest for your very useful recommendation,
you said to the great, all the wisdom
you do not know what exactly was running (what services) but must have stupid comment
all other people who also report performance issues / degraded working node here are probably also wrong..
and as i mentioned at all other VPS at other (reliable and stable) providers i have no issues w the same my "shits" and the same config never and ever, it's miracle..
and simultaneously you are sure greater expert as people who develop Virtualmin, because they clams recommended minimum ram for minimal installation package is 256MB, i'm glad you have led them out of mistake now
@miu do you have any other 256MB KVM VPS that you have tried this on?
yes, several (KVM and also oVZ), and if u border self read whole this thread u can find it above - screen of latest. I can guarantee from own experiences that config from 256 ram & swap works without issues, stable, longterm (minimal package w spamassassin and clamav disabled), if node is fast and stable (not overloaded, not degraded, or not oversold)
ya sure you did.
Looks like someone managed to tackle LET stuff, instead of petty fighting but with solutions...
And then I came to see @Cloudcone is an old member
.... BTW can you not limit IOPS (so it does not affect other users of the node) or Add a IOPS graph for users...
no-possible "shit" especially captured for you, running on stable and fast SSD KVM node:
Used RAM: 139 MB of 236 MB Total
(used swap 302 MB)
Uptime: 227 days
CPU load aver. (15 min period): 0.01
227 days running these services absolutely without one issue and restart:
Apache, BIND, MySQL, postfix, dovecot, webmin service, firewalld, fail2ban, SSH, crond
and second screen after another half hour later, as first one was made:
honestly you should not be hosting virtualadmin with your mail progs/mysql/etc on 256mb ram. The issue is the swap space and how each provider handles it.
Your graphs are meaningless because each provider can determine abuse in their own way because once swap is used; the iowait can cause cpu to spike more than 10x normal.
you can see also 512 MB swap would be no problem and still working w reserve
purpose was show that discussed config in this thread 256 MB + 1000 MB swap can be fully enough , simultaneously CPU load is still minimal.. why?? because HOST NODE IS really RAPID & FAST
additionally i think bigger swap (at small ram 256 - 512 MB) is no problem or possible threat for high CPU load when node is fast SSD w good IOPS and you simultaneously do not do great load (mean run only 1-2 no-robust website w emails what are not difficult on resources, as in my case)
i think all depend how stable fast and crowded is concrete host node (provider quality and their definitions and standards for servers performance and load) and if over-sellers placed tons of vps instances on one poor node then nothing cant be working well and smooth there (woot, alpha, hiform, nfp etc..)
and of course i agree that when you plan host 50 or 100 dynamic websites w regular MySQL using i think then need any such 4 or 8 MB RAM and several CPU cores
Although I too think 256 MB are definitely too few for all of those daemons running, you have a point here. When a system is fast, the operations complete quickly, thus freeing up memory. When a system is slow, more operations tend to be executed simultaneously, requiring more memory at the same time.
But this doesn't mean that Cloudcone is to blame here, you just compared two different products (maybe at different prices?).
May I ask which provider this server is from?
@v3ng
check PM