Fine-tuning diskless clients, lagg & monitoring

[Beeblebrox]

1. Is there a monitoring tool (GUI or console based) for diskless clients (load, usage, etc). Is there a way to remotely poweroff the clients WITHOUT using ssh or telnet?

2. Is there a reason why distcc settings in diskless/etc/rc.conf fail to start and need to be started manually?

distccd_enable="YES"
distccd_flags="--daemon --allow 192.168.2.1"

3. lagg setup: Instructions describe modifying /etc/rc.conf for lagg device. This seems an odd solution to me in diskless environment. Is it possible to pass the lagg parameter for the second NIC through the dhcp server? Target mode is loadbalance. This setting in diskless/etc/rc.conf freezes the system (re0 is already up). I suppose the nfs-mounted root on diskless becomes confused by the re0 -> lagg0 switch and looses the mountpoint. Why do I want lagg on a diskless client? the mobo-embedded NIC is 10/100, wile the second NIC card is Gbit. My switch is unmanaged.

ifconfig_re1="up"
cloned_interfaces="lagg0"
ifconfig_lagg0="laggproto loadbalance laggport re0 laggport re1 192.168.2.2 netmask 255.255.255.0"

4. A diskless node which has no gpu fails to boot (does not respond to ping) but boots fine when I place a gpu card in the PCIe slot. What should I disable (xorg?) in rc.conf for this? I can't get to error logs as /var is mounted as md0.

 

[Beeblebrox]

Any attempt to configure lagg0 on diskless client (DC) results in immediate lock-up of the NIC's and causes system freeze. The DC no longer responds to pings after
# ifconfig lagg0 laggproto lacp laggport re0 laggport re1 192.168.2.1 netmask 255.255.255.0

 

[Beeblebrox]

Problem 3 now seems to be concerning lagg its self. I have two NIC's on the main system also and I tried pooling them, both as lacp and as loadbalance.
# ifconfig lagg0 laggproto loadbalance laggport alc0 laggport re0 192.168.1.10 netmask 255.255.255.0
Then trying to ping the gateway gave:

PING 192.168.1.1 (192.168.1.1): 56 data bytes
ping: sendto: Device busy
ping: sendto: Device busy

So obviously that's what happens to the diskless machines as well - the NIC pool becomes "busy" so the system freezes as it can't communicate with master. Do I have to get down to the details of ng_one2many to get this to work?