| Author |
Message |
coolDude
Joined: 29 Jan 2004
Posts: 3
|
| Posted: Tue Feb 17, 2004 9:00 am Post subject: Memory limites |
|
|
Hello!
If I have a Linode 96, does it mean that the maximum amount of RAM available to me is 96MB?
I've heard that in my case(Linode96), 96MB would be the minimum guaranted RAM for my Linode, but in case there is free memory available on the server, I may be using more than 96MB. Is this true? :roll:
Thanks! :) |
|
| Back to top |
|
blahrus
Joined: 19 Jan 2004
Posts: 35
Location: Bloomington, IL
|
| Posted: Tue Feb 17, 2004 9:13 am Post subject: |
|
|
Nope that IS the amount of RAM you are allowed . . . . Just make sure you give your self at lest 256 megs of swap and you should be fine.
(Chris kick me if I am worng) |
|
| Back to top |
|
adamgent
Joined: 23 Jun 2003
Posts: 261
|
| Posted: Tue Feb 17, 2004 9:30 am Post subject: |
|
|
Hi,
blahrus is right you are limited to the amount of ram for your linode.
The only thing that is burstable is the CPU usage, if there is spare CPU time you will use it if you need to.
Adam |
|
| Back to top |
|
blahrus
Joined: 19 Jan 2004
Posts: 35
Location: Bloomington, IL
|
| Posted: Tue Feb 17, 2004 9:43 am Post subject: |
|
|
so no one on host12 is allowed to use gentoo :)
jk! |
|
| Back to top |
|
thyrinn
Joined: 13 Feb 2004
Posts: 18
|
| Posted: Tue Feb 17, 2004 4:45 pm Post subject: |
|
|
blahrus wrote: so no one on host12 is allowed to use gentoo :)
jk!
Exactly, I need it for running my seti@home and distributed.net j/k :)
How does the CPU scheduler work for UMLs? Do all the UML processes run at the same priority on the host and all CPU is split even? If two Linodes both try and use 100% CPU they each get 50%? |
|
| Back to top |
|
caker
Joined: 15 Apr 2003
Posts: 2392
Location: Galloway, NJ
|
| Posted: Tue Feb 17, 2004 4:47 pm Post subject: |
|
|
thyrinn wrote: How does the CPU scheduler work for UMLs? Do all the UML processes run at the same priority on the host and all CPU is split even? If two Linodes both try and use 100% CPU they each get 50%?
The CPU scheduler on the HOST distributes CPU time evenly across all the processes (UMLs) asking for CPU time... That is why I group Linodes of the same plan on one box. So yes, if two Linodes want full steam, they'll both get 50% (actually, they'll both get 100%, since these are dual-processor machines).
-Chris |
|
| Back to top |
|
thyrinn
Joined: 13 Feb 2004
Posts: 18
|
| Posted: Tue Feb 17, 2004 4:47 pm Post subject: |
|
|
| UML is supposed to run as 4 processes (I think) with the skas patch, and the hosts are SMP machines, so how does all this relate performance wise? Can one linode max out both CPUs? |
|
| Back to top |
|
caker
Joined: 15 Apr 2003
Posts: 2392
Location: Galloway, NJ
|
| Posted: Tue Feb 17, 2004 4:50 pm Post subject: |
|
|
thyrinn wrote: UML is supposed to run as 4 processes (I think) with the skas patch, and the hosts are SMP machines, so how does all this relate performance wise? Can one linode max out both CPUs?
The two primary threads of an SKAS UML are the Kernel and User thread. Since UML switches between the two, technically the UML is single-threaded. The other two threads are an async-IO thread and a signal delivery thread. I have seen UML utilize > 100% of one CPU indicating that it is gettingsome benefit from SMP.
-Chris |
|
| Back to top |
|
| |
