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 thought so too but there were other providers that had the same issue, some fixed it by moving away from SolusVM but others just accepted it. We removed our plans under 128MB to keep it above 100Mbps.
And my effort above is completely overlooked @KuJoe...
Well the netspeeds were hit much harder on the old 3.x.x Versions and TC certainly does not help.
I was one of those providers, imho its not SolusVM, its just that SolusVM by default puts old versions of stuff on your node, upgrading is essential.
Its all proportional obviously but even the 64mbit plans can get over 300 mbit now.
It wasn't overlooked. I noticed it and was going to comment on it but Anthony replied with a possible explanation for my test results so I've decided to look into our setup to try to achieve the same results you posted. Sorry for not acknowledging your post but I didn't know what to say about it except to use the data you provided as reference.
@KuJoe I was just kidding. Already morning here...
Any better results with advice from Anthony?
It actually is a OpenVZ vps, so that might make a difference.
So we would probably need like 96mb on XenPV to achieve the same i thik.
Thanks for informing me about the heavy disk I/O though, didnt know that!
Might share the install script for it
pretty handy!
@Spirit, those sorts of changes require node reboots and a day or so to play with so I would not expect a quick turn around
@djvdrop, be kind to your host and find the preallocation flag for rtorrent and reduce your IO
I use it as my OpenStatus server
(edit) Nevermind, LET loves to start me on page 1 for some reason. (/edit)
is the RAM sufficient for running a webserver with OpenStatus etc on it?
Nginx probably?
If it is, might be interested to use it as openstatus monitor maybe
PS: just like your datacentre @AnthonySmith, I am also located to the south of Rotterdam, NL
@djvdorp it's running apache
@birdie25, I am stunned running openstatus on apache that it is not sitting in swap by now, you might want to check that?
I will pull an IO chart and see if any of the 64mb plans are abusing IO yet..... prepare for some wrath! hehe
@AnthonySmith It's in swap with 10MB
Well I guess someone lied a little when they pressed the "I know the SWAP is not a RAM substitute! box"
http://i.imgur.com/J3c03.png
That's from 4 of them on Node 2, charting the other nodes now.
This is exactly the reason more hosts do not offer these.
@birdie25 sorry that post was just bad timing I was not taking to you I have not matched VPS's to names yet
Very sad. Maybe it would be possible to automatically monitor the swap usage and pause a VPS if it is consuming swap for more than 1 hour or within any other specific timeframe?
@AnthonySmith:
What's about, that potential customers for the 64mb-plan must answer some linux / unix questions?
@nabo I understand the idea but unless its serious abuse I just ask people to fix the issue with a 24 hour deadline support ticket that way everyone co-operates and it helps people re-align their expectations
I have been working on an auto alerting system for months and keep starting again from scratch as I am never happy enough to put it in production, if anyone is a bit handy with RRD drop me a PM there may be a free VPS in it for them
I warn you though the available output is horribly formatted
@Amfy again I like the idea but the problem with that is the human element+google they will just lie I expect
Not the end of the world though these things are not hard to manage, what I find very hard to manage on occasion is peoples expectations,,, I get lots of answers like this even when presented with evidence "That cant be correct.... I have 10 VPS's running the same thing and no host has ever had a problem with it"
Which to me just says.... "My other hosts don't bother monitoring anything so I get away with it thinking everything is fine while other people sharing the node have a hard time"
@birdie25 Linux likes to swap even with multiple GBs of free RAM.
Put your swappiness down ?
My current usage is
# free -m total used free shared buffers cached Mem: 64 54 9 0 6 35 -/+ buffers/cache: 12 51 Swap: 63 0 63@AnthonySmith: but it would be a try worth, you could do that only on the small packages.
hehe, or ordering with a telnet-similar client? :P
Yeah, I understand. That's the point where I can't understand lowend-providers and having my full respect. A 2-3$ customer is going on their nerve's all the time and want the whole node for themselve.
Or they want big support with install helping for an small unmanaged plan.
And then everyone is crying when providers start massive overselling :P
And that all is the reason why I only sell a few webspace/shell and managed services (some small loadbalanced things) where the customers get absolute no root access!
free -m total used free shared buffers cached Mem: 58 56 1 0 8 17 -/+ buffers/cache: 30 27 Swap: 63 11 52And that's normal. Some pages are never needed and it's better to have them in swap.
All the trouble comes when someone thinks that he has 128 MB RAM when he has 64 MB RAM and 64 MB swap.
I was thinking - maybe make these boxes swapless? Or give them only small ammount of swap - 10-15 MB? Linux will be happy and user will not think "wow, I have 2x more RAM for the same price, lets use it".
Because frankly I don't see where LEB-applications could have unused pages of memory.
@breton: Good idea. And users which are really needing, for normal reasons a higher amount of swap and have the technical know-how will be able to add swap at themselves
name a couple here, please.
Or maybe somebody could name them?
updates often need more than 64mb ram
Yes, I also would say updates. Or when you run many services, but you don't need them often. For example munin-processes.
But I never used swap much, so not sure.
If you are worried about using swap just use swapoff to disable it.
HA, switched apache to nginx, reduced my swap!
free -m total used free shared buffers cached Mem: 58 56 1 0 12 23 -/+ buffers/cache: 21 36 Swap: 63 7 56