vmware a "utekajuci cas" ..
Jan Rusnak
jx06 na felis.sk
Středa Srpen 9 09:13:44 CEST 2006
Kozo wrote:
> Zdravim ..
>
> stretol som sa so zaujimvou skutocnostou a chcem sa poradit ..
>
> Vsimol som si, ze na testovacich serveroch, ktore mam nainstalovane vo
> vmware (CentOS 4.x) mi "uteka" cas. Nastavim ho cez ntpdate, spustim
> ntp daemona a o pol dna mam (napriek beziacemu ntpd) cas posunuty radovo
> o hodiny .. v step-tickers mam zadefinovany pool.ntp.org a v ntpd.conf
> zase standardnu trojicu:
> server 0.pool.ntp.org
> server 1.pool.ntp.org
> server 2.pool.ntp.org
Na virt. strojoch nepouzivajte sync. casu cez ntpd, to z principu
nefunguje. Mozete pouzit sync. cez vmware-tools, ako to uz tu niekto
pisal. Vmware-tools ale "posuva" cas iba jednym smerom, teda zabezpecuje
dobiehanie realneho casu.
Cas na VM mozu ovplyvnit nastavenia guest kernelu "Timer frequency"
(100Hz) a Kernel timer.
Precitajte si toto:
Option clock=pit is the simplest and most robust of the available
algorithms. It uses only the PIT counter for interpolation. This option
does not include lost tick correction code, so it does not gain time,
but it does lose time when a tick is actually lost. This option
is recommended for use in VMware virtual machines, together with VMware
Tools time synchronization, to correct for the occasional lost ticks.
--
JR
----------------------
JAN RUSNAK
<jx06 na felis.sk>
phone: +421 903 550345
----------------------
Další informace o konferenci Linux