Re: [zones-discuss] Re: Java VM defaults in memory capped zones

2007-05-18 Thread Jason Bradfield
Thanks Steffen, The memory capping from the global zone is why we decided to go with Nevada instead of Solaris 10... We are desperately waiting for dedicated memory, Jeff said this is not on the cards for 2007 though... The problem we are finding with capping memory is that our clients and m

Re: [zones-discuss] Re: Java VM defaults in memory capped zones

2007-05-18 Thread Steffen Weiberle
Hi Jason, coming late into this thread, Jason Bradfield wrote: Thanks Jeff. FYI, These zones will be allocated to different clients each with root access to their zone and running their own software. Whether you choose to provide root access to the non-global zone users has no effect on the

Re: [zones-discuss] Re: Java VM defaults in memory capped zones

2007-05-17 Thread Jeff Victor
Jason Bradfield wrote: Thanks Jeff. FYI, These zones will be allocated to different clients each with root access to their zone and running their own software. Ah. Yes, that would indicate tighter controls. You could even charge more for a higher cap... ;-) Jeff Victor wrote: Jason B

Re: [zones-discuss] Re: Java VM defaults in memory capped zones

2007-05-17 Thread Jason Bradfield
Thanks Jeff. FYI, These zones will be allocated to different clients each with root access to their zone and running their own software. What we really want and are waiting for is dedicated memory, not capping.. Do you know when this will be coming, and will this also include prstat, df, swap

Re: [zones-discuss] Re: Java VM defaults in memory capped zones

2007-05-17 Thread Jeff Victor
Jason Bradfield wrote: Sorry if this comes through as a duplicate response but I responded from a different email account earlier... Thanks Bob and Jeff... Therefore here is my confusion.. From Bob * swap: the amount of disk space set aside to copy physical memory to when copying it out of R

Re: [zones-discuss] Re: Java VM defaults in memory capped zones

2007-05-17 Thread Jeff Victor
Jason Bradfield wrote: Actually put simply how can both of these be true... From Bob * swap: the amount of disk space set aside to copy physical memory to when copying it out of RAM From Jeff set swap= this is VM - the total address space available to the zone Bob and I used differe

Re: [zones-discuss] Re: Java VM defaults in memory capped zones

2007-05-16 Thread Bob Bownes
>> Can anyone, in lamens (ex linux) terms explain the differences between >> virtual/swap/rss/size/physical etc or recomend a good >> blog/article/document. > >I don't know Linux terms, so I'll define terms for this conversation: >* physical memory: RAM, memory chips >* VM (virtual memory): the tot

Re: [zones-discuss] Re: Java VM defaults in memory capped zones

2007-05-16 Thread Jeff Victor
Jason Bradfield wrote: Also, on a similar note can someone confirm my understanding of the physical and swap caps... In my case above will that give the zone a total of 256M Virtual memory of which no more than 128M will be physical.. My tests seem to prove this.. Yes, although the zone is all

[zones-discuss] Re: Java VM defaults in memory capped zones

2007-05-16 Thread Jason Bradfield
Also, on a similar note can someone confirm my understanding of the physical and swap caps... In my case above will that give the zone a total of 256M Virtual memory of which no more than 128M will be physical.. My tests seem to prove this.. My initial thoughts were that I would have 384M of wh