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
@eastonch master is only hosting one of my text vps. No clients on master. So no, not a bad practice.
@Damian I think it is the 57 stable. I will check back once I get off from my bed.
We recently had issues with .32 and KVM, as was mentioned before this was nothing to do with cache and no RAM would free up. A 16GB node with only 4 256 guests was using like 14GB of RAM and the only way it would free up was restarting, then the leak would build gradually at rates of 256MB every 15 minutes.
It ended up been related to having traffic shaping enabled, although i don't know if this is the same problem your having.
So yeah, we now have IP stealing disabled as it stops IPv6 working and traffic shaping disabled to stop memory leaks. I would love to use all the features SolusVM offers one of these days :P
@NinjaHawk still find hosting any sort of VPS, or any "unneeded" load on a master is silly.
The master literally, should be a control panel, sat there, secured to fuck. On something like a Linode, or a small dedicated server. Automatic backups every few hours. It's how mission critical I find it. :']
If i had my way, there wouldn't be a master. IE no possible central security breach.
@GetKVM_ash this is what my original plan was with Solus. But later, everything went the other way when I realised that it would be a nightmare with solusvm configuring each node+whmcs everytime you get a new node.
Pfff with this threads....
And you have a hosting company...
@yomero what do you mean by that? Obviously I am not the only one.
Whatever
Thanks for the clarification.
This is a shame, period.
@Zen I know how memory is managed, this isn't it. People who just assume that everyone reporting this issue is an idiot who doesn't understand memory management are a problem in this case. You've got mixed messages going out, "experienced" sys admins being lazy and assuming the kernel is unstable and that they can completely ignore an out of control cache situation because "lol I can read linuxatemyram.com."
This is not a normal cache operation, it causes instability, and it is a bug.
@ZEn Your quote doesn't make sense, what I was referring to was the fact that 2 node crashed without any warning and it turned out that the memory was full and cache was not being released.
Call me dumb but when something doesn't work the way it was supposed to, I will ask.
No "you guys" here. I understand caching. Snowballing until the kernel crashes is not even possibly a bug, it IS a bug. Not selling yourself as a freelance sys admin well if you just go "lol it's cache" and assume you're not the idiot. Ninja was honestly asking if someone was suggesting that they had knowledge that this cache scenario was fine and would release the memory. Valid question, since it won't in this case. I hope you aren't a "provider" @Zen (since we're being dicks about it, count me in, I'm pretty good at that game)