I have 5 AMD Athlon 1.6Ghz with 1Gb RAM. plenty of swap, /boot,/,etc.
The machines had only bind 9.x running and serving a fake domain (lab01.com, lab02.com, lab03.com,lab04.com,lab05.com) with internal IP addresses.
Each machine responds and resolvs correctly to his own domain.
Well, this setup was initiated at 4pm yesterday and is still running on some machines. Why does it take so long?
I use Centos 4.4 stock, minimum install, dependencies for zimbra were provided and the machine is clean of any other software.
I'm using the 4.0.5 release for Redhat 4.4 RPMs.
I have also noticed an incredible amount of swap usage, up to 2gigabytes of swap have been used at a given time.
The myslq daemon sometimes goes all the way to 900megabytes of ram usage.
In the only machine that the software finished the installation I did a reboot and started zimbra. zmcontrol status takes forever to show something.
Funny thing: All the machines had almost 10k sleeping processes during zimbra setup, then, i switch to a virtual console and started TOP. Using top I noticed that suddenly the sleeping processes were decreasing and at the same time the zimbra installation started to "install" again.
Is my machine entering some unknown sleep process? or the hard disk?
I did not enable anything related to apm in my centos.