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
Hahaha! Sorry, I need to abuse
Let me do it please 
!!!
Go for it
Francisco
With a 20 load Fran will suspend me
If this was production, sure, but it's a test bed for a reason
Francisco
Can someone explain why am I hitting the privvmpages limit here?
# free total used free shared buffers cached Mem: 524288 336540 187748 0 0 4156 -/+ buffers/cache: 332384 191904 Swap: 0 0 0 top - 13:43:53 up 2 days, 2:32, 2 users, load average: 38.00, 37.64, 29.60 Tasks: 44 total, 1 running, 43 sleeping, 0 stopped, 0 zombie Cpu(s): 0.0%us, 0.6%sy, 79.5%ni, 19.9%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 524288k total, 337936k used, 186352k free, 0k buffers Swap: 0k total, 0k used, 0k free, 4224k cached # cat /proc/user_beancounters Version: 2.5 uid resource held maxheld barrier limit failcnt 15059: kmemsize 12029609 24764416 2147483646 2147483646 0 privvmpages 261781 262144 262144 262144 9367 physpages 82451 131559 131072 131072 10 oomguarpages 78468 128638 131072 2147483647 3There is plenty of free RAM left, but I can't allocate anything else. If looks similar to what @dmmcintyre3 posted, but my physpages has a decent margin left.
Hah, @yomero, I found your stress tests :P
Well done.
The nodes sitting at 35 loads since you're not the only one stressing it.
Francisco
@vedran - that's just an old setting since stallion isn't 100% vswap/.32 aware.
I can set you to 'unlimited' if you PM me your IP.
Francisco
Yeah, I guess that all of us have the same problem
With my apps I found my barrier around 600MB (1GB box).
Just having fun lol. Never stressed so hard a VPS :P In particular my VPSs with you have less than 1.0 load u_u
Funny thing is, I can allocate all 512MB if I do it from one process, but multiple processes are apparently using more than it's reported by free. Or privvmpages measures memory usage the way "old" OpenVZ did. Or I have no idea what I'm talking about :S I really can't understand how the new user_beancounters works.
Either way, I can't crash this node no matter how hard I try. It's very responsive even with 33.97 load I'm having right now. But I'm new to this abusingload testing field, give me some time.
lol!
Well, I've read that ubc is being discontinued, so, is unneccesary to establish the memory limits. But still some of these counters influence in the system behavior.
Oh come on @yomera surely you can do better than 32.
What are you? Chicken?
yomera is like... female u_u
And yes, I can raise more and more this loads, but Fran will get crazy XD
bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk bwauk
(edit: That's supposed to be chicken clucking there if you haven't guessed.)
Yes, I guessed ¬_¬!!
Mine hasn't arrived yet
Got an order ID for me?
@yomero - I'm actually waiting to see where the breaking point is to be honest. I mean, 40 loads is bad but people are reporting that node99 feels very responsive so I dunno. The load is there since I see the amount of CPU that's tanked :P
Francisco
Well, I guess that is because the CPU scheduling, and then I am not interfering with the other people, despite the increasing process queue :P
Well, it's handling CPU scheduling much better either way.
Francisco
Yup, I do, the magic number is 9401438679.
I see it, thanks Fran
It looks like your VPS got auto provisioned during one of the box crashes that happened.
Sorry about that!
Francisco
@Francisco Np
We've restocked our trial VM's:
https://my.frantech.ca/cart.php?a=add&pid=99&carttpl=cart https://my.frantech.ca/cart.php?a=add&pid=100&carttpl=cart
We nuked everyone that was sitting at a 0.0 load. Abuse it or lose it people!
Francisco
Was nuking it for a while, but got tired of being locked out
The lockups have been addressed so far
There's 2 bugs we know of that we're rolling up ksplices for and i think we'll be sitting good.
Francisco
To poke fun at @yomero, Anthony was doing the mass load check to see who to hit. He found yomeros sitting at a ~30 load and goes 'theres a keeper!'
Francisco
I compile kernel. Load is now about 8 and it going to up.
maybe some suggestions besides the compiling kernel would be helpful...
It seems that OpenVZ is not best platform compile kernel, so I try another way generate load.
When I calculate floating points, load average is over 65. If I run two or more floating points calculation in same time, I can get higher loads. Load level is now about 130, when I run two floating points calculation in same time.
And if I run two floating points calculation in same time with different values, I got about 257.
HAHAHAHA LOOOOOOOOOOOL!!!!
I am the winner XD
I wanted to go up to 50 load or more but, I can't allocate all my memory u_u
Btw, do you rebooted the node? Crashed? Now I have 3 hours of uptime
Edit: tux is crazy u_u