Re: [ilugd] Limiting physical memory & observing consequences...

2005-11-18 Thread Mayank Jain
On 11/18/05, Tushar Shah <[EMAIL PROTECTED]> wrote: > Hi, > You can create tmpfs and fill it and then observe the effects on > processes running out of mem . > > http://www-128.ibm.com/developerworks/library/l-fs3.html Probably this link will interest many... http://www.cs.berkeley.edu/~nks/fi

Re: [ilugd] Limiting physical memory & observing consequences...

2005-11-17 Thread Tushar Shah
Hi, You can create tmpfs and fill it and then observe the effects on processes running out of mem . http://www-128.ibm.com/developerworks/library/l-fs3.html Regards Tushar Shah -- Somedays you're the dog, and some day's you're the hydrant ___ ilugd

Re: [ilugd] Limiting physical memory & observing consequences...

2005-11-17 Thread Sirtaj Singh Kang
On Thursday 17 November 2005 22:17, Mayank Jain wrote: [snip] > What about doing this on per-application basis. > The above method would limit every app to the specified number... > > Say, I want to limit Open Office to 100megs...? You can try playing with the ulimit command, or coerce a memory p

Re: [ilugd] Limiting physical memory & observing consequences...

2005-11-17 Thread Mayank Jain
On 11/17/05, Sirtaj Singh Kang <[EMAIL PROTECTED]> wrote: > By default the kernel also overcommits on memory allocations by apps. To turn > this off this should probably work: > > echo 2 > /proc/sys/vm/overcommit_memory > echo 100 > /proc/sys/vm/overcommit_ratio > > (not 100% certain about this, pl

Re: [ilugd] Limiting physical memory & observing consequences...

2005-11-17 Thread Sirtaj Singh Kang
On Thursday 17 November 2005 18:16, Mithun Bhattacharya wrote: [snip] > Once it runs out of physical memory new request for memory would result > in swap usage. If you manage to use both the physical memory and the > SWAP the OS will return a error. By default the kernel also overcommits on memor

Re: [ilugd] Limiting physical memory & observing consequences...

2005-11-17 Thread Abhay Kedia
On 11/17/05, Mayank Jain <[EMAIL PROTECTED]> wrote: > > I wonder how apps like X, GDM, KDE/Gnome, etc would react to > "insufficient memory" problems... > > Like if before starting these heavy services/apps, if we could limit > the physical memory to say... "100 megs" on a system & then see how do

Re: [ilugd] Limiting physical memory & observing consequences...

2005-11-17 Thread Mithun Bhattacharya
--- Tushar Shah <[EMAIL PROTECTED]> wrote: > Hi, > > > & specially, it would be interesting to see how they react to low > mem > > conditions, do they crash or exit gracefully or something else... > Why shud it crash ? I guess such a situation would cause lots of > page > swapping . Better a

Re: [ilugd] Limiting physical memory & observing consequences...

2005-11-17 Thread Tushar Shah
Hi, > & specially, it would be interesting to see how they react to low mem > conditions, do they crash or exit gracefully or something else... Why shud it crash ? I guess such a situation would cause lots of page swapping . Better answer can come from some who knows abt linux memory management

[ilugd] Limiting physical memory & observing consequences...

2005-11-17 Thread Mayank Jain
Hi :-) Was just replying to Ram's message when this idea struck... I wonder how apps like X, GDM, KDE/Gnome, etc would react to "insufficient memory" problems... Like if before starting these heavy services/apps, if we could limit the physical memory to say... "100 megs" on a system & then see h