hi guys,
just want to share my recent experience about upgrading my freenas box to 10G, below included my issues and workaround. Hope someone could provide some tips and fix.
Basic environment:
one server will freeze during the last stage of the boot. (ping success before the freeze, lost ping after freeze), however, the other one no issues.
Current configurations on the switch side is: lacp based port channel (mode active) (for both servers, each server has two twinax cables connected to different switches (one port for one switch), (switch side: two ports formed a port-channel)
workaround: with my co-worker's help, added 10seconds delay in the dbus daemon fixed this issue,
thoughts: looks the cards are having issue talking to switch to form the port-channel (need a few seconds to finish) (inside single user mode, nic is working fine)
Issue 2:
followed issue1, I tested the failover for the two links, disable one port in the port channel, the network was broken, the other link was not brought up. see attached screenshot, one interface in the lagg is not showing status.
workaround: have to change the lagg configuration from lacp to failover and everything is functioning as expected.
Steps done: swapping NICs between servers did not make any difference, swapping ports on switch side did not make any difference. changing card slot in the issueed server did not make difference.
Although two servers are working fine right now, still hope the LACP will work.
just want to share my recent experience about upgrading my freenas box to 10G, below included my issues and workaround. Hope someone could provide some tips and fix.
Basic environment:
- two iXsystems certified freenas servers (32G RAM, two cpus, 12X6T sata drives)
- freenas version: 9.10 stable 201604140622 (8781269)
- two Dell intel X520-DA2 ethernet 10G dual port
- two Dell N4032F stacked together
one server will freeze during the last stage of the boot. (ping success before the freeze, lost ping after freeze), however, the other one no issues.
Current configurations on the switch side is: lacp based port channel (mode active) (for both servers, each server has two twinax cables connected to different switches (one port for one switch), (switch side: two ports formed a port-channel)
workaround: with my co-worker's help, added 10seconds delay in the dbus daemon fixed this issue,
thoughts: looks the cards are having issue talking to switch to form the port-channel (need a few seconds to finish) (inside single user mode, nic is working fine)
Issue 2:
followed issue1, I tested the failover for the two links, disable one port in the port channel, the network was broken, the other link was not brought up. see attached screenshot, one interface in the lagg is not showing status.
workaround: have to change the lagg configuration from lacp to failover and everything is functioning as expected.
Steps done: swapping NICs between servers did not make any difference, swapping ports on switch side did not make any difference. changing card slot in the issueed server did not make difference.
Although two servers are working fine right now, still hope the LACP will work.
