Astaro 6.311
9 Interfaces, eth0 builtin, eth1-4 and eth5-8 on 4-port cards.
eth0 is 'public' interface
Other interfaces are a mix of publicly and privately addressed sub-nets with different masks.
DHCP offered on MOST but not all interfaces.
At one point I had DHCP configured for eth0, my public net. It never worked right,it seemed, and I really did not want it anymore.
I tried to set all the values to nothing and save, but that was unacceptable. So I noticed that I could disable DHCP by interface (Dah), so I just disabled it for eth0.
Some time later I noticed that eth2 (192.168.128.0/28) was not serving DHCP leases, and even statically addressed devices were not pingable and could not ping even the firewall interface. Interestingly enough, I saw the traffic when I did a ifconfig eth2. Data was moving but nothing happening.
I thought maybe eth2 had gone bad on me, so I set eth7 up with these addresses (moving eth2 to a different net). Since all the interfaces are plugged into the same switch, I just switched which VLAN the interfaces were on and PRESTO. It still did not work. Now it was 2am in the morning and I was getting desperate.
I noticed that with eth2, I just set it up for 192.168.64.0/28, so I moved it back to its original VLAN (moving eth7 back to its VLAN), statically addressed some devices on that VLAN and PRESTO they started working. So I quickly statically addressed all the important stuff and went to bed.
Now I am over-tired, and could not get DHCP leases for devices that wanted them. Well on some nets. eth8 was not having any DHCP problems. I thought there might be something in the configs messing up 192.168.128.1. I SSHed in, SUed, and greped everything. Did not see anything that could point to a problem. Then I thought back that I made a change to my DNS internal view, adding these private addresses, but not doing the reverse, yet. But again, inconsistant. Not DNS. ?????
Out of frustration, I enabled DHCP on eth0, setting it up for only 2 addresses (why can't I specify the same start and stop address for only one DHCP lease address?).
AND PRESTO IT WORKS AGAIN.
Well later I will put all the addressing back, but this begs the question.
WHAT WENT WRONG?
It seems that disabling DHCP on an interface that was serving leases had caused DHCP to go haywire.
Oh, I did reboot the system in all of this, thinking that MAYBE my problem would be 'fixed' by a reboot. Nope.
It looks like I am operational, but I REALLY do not want DHCP leasing on eth0....
Please help. I need to go to slllleeeeepppppp.
This thread was automatically locked due to age.