August 1, 20169 yr I have a Dell PowerConnect 2724 Switch that supports 802.3ad LACP and I've used it for a while on my Mac. I run unRAID on an ASUS P9D-E/4L with 4 Gigabit LAN ports. I created a new LACP group on the switch, connected all the cables, but as soon as I tell it to use 802.3ad in the network settings it brings down ALL the LACP groups(not just the unRAID ones), and doesn't connect to the network causing a big mess. Any help?
August 1, 20169 yr can you post some syslog output when the changes goes active? What is your Min/Max members per lacp group? what other groups are configured for lacp? what is the load balancing algo for your lacp setup?
August 1, 20169 yr And some more background information about your unRAID system. What version is it running and post a screenshot of your network settings.
August 2, 20169 yr Author And some more background information about your unRAID system. What version is it running and post a screenshot of your network settings. These are my network settings except for the bonding type(if I change it to 802.3ad it breaks) I'm running 6.1.9 Trial (for now)
August 2, 20169 yr Author can you post some syslog output when the changes goes active? What is your Min/Max members per lacp group? what other groups are configured for lacp? what is the load balancing algo for your lacp setup? I'm a total noob, how do I post the syslog output? On the switch I can create up to 6 groups(I'm using 2, 1 for unRAID and 2 for my Mac) of up to 8 ports(I'm using 2 ports for the Mac and 4 for unRAID). The load balancing I'm using is 802.3ad which is supported by the switch.
August 2, 20169 yr Take some screenshots of the switch lacp config, there are different hashs for loadbalancing with multi-links, mac, mac + ip, ip+port, etc etc. Im interested to see what your min/max links are for this new lag, im assuming you set min/max at 4. When you switch the config on, can you look at your servers output of /proc/net/bonding/bond0 and post the output, interested to see if the heartbeat is making it on to your nics. Tim can you post some syslog output when the changes goes active? What is your Min/Max members per lacp group? what other groups are configured for lacp? what is the load balancing algo for your lacp setup? I'm a total noob, how do I post the syslog output? On the switch I can create up to 6 groups(I'm using 2, 1 for unRAID and 2 for my Mac) of up to 8 ports(I'm using 2 ports for the Mac and 4 for unRAID). The load balancing I'm using is 802.3ad which is supported by the switch.
August 2, 20169 yr Author Take some screenshots of the switch lacp config, there are different hashs for loadbalancing with multi-links, mac, mac + ip, ip+port, etc etc. Im interested to see what your min/max links are for this new lag, im assuming you set min/max at 4. When you switch the config on, can you look at your servers output of /proc/net/bonding/bond0 and post the output, interested to see if the heartbeat is making it on to your nics. Tim This is my switch's configuration, it seems to be purely port-based multi-link. I can't post the output because as soon as i click Apply it just goes down and I can't navigate the webgui. To bring it back up I had to wipe the USB drive and reinstall everything.
August 2, 20169 yr Take some screenshots of the switch lacp config, there are different hashs for loadbalancing with multi-links, mac, mac + ip, ip+port, etc etc. Im interested to see what your min/max links are for this new lag, im assuming you set min/max at 4. When you switch the config on, can you look at your servers output of /proc/net/bonding/bond0 and post the output, interested to see if the heartbeat is making it on to your nics. Tim This is my switch's configuration, it seems to be purely port-based multi-link. I can't post the output because as soon as i click Apply it just goes down and I can't navigate the webgui. To bring it back up I had to wipe the USB drive and reinstall everything. It sounds like your switch and the unRAID system don't understand each other, and hence a layer 2 loop is created. Start with connecting just ONE port and then apply the setting. You may need a console connection to verify the state of the ethernet port on the unRAID system.
August 2, 20169 yr Your switch does not support dynamic LACP, it is another broken dell implementation. Id recommend the Cisco SG-200 series if you want a mondernish switch that supports most things. Ive upgraded to D-Link 10GB in the house because LACP is rather crappy. http://en.community.dell.com/support-forums/network-switches/f/866/p/19169209/19297716 Static LAG is your only option, if you loose one link your hosed. Tim Take some screenshots of the switch lacp config, there are different hashs for loadbalancing with multi-links, mac, mac + ip, ip+port, etc etc. Im interested to see what your min/max links are for this new lag, im assuming you set min/max at 4. When you switch the config on, can you look at your servers output of /proc/net/bonding/bond0 and post the output, interested to see if the heartbeat is making it on to your nics. Tim This is my switch's configuration, it seems to be purely port-based multi-link. I can't post the output because as soon as i click Apply it just goes down and I can't navigate the webgui. To bring it back up I had to wipe the USB drive and reinstall everything.
August 2, 20169 yr Author Take some screenshots of the switch lacp config, there are different hashs for loadbalancing with multi-links, mac, mac + ip, ip+port, etc etc. Im interested to see what your min/max links are for this new lag, im assuming you set min/max at 4. When you switch the config on, can you look at your servers output of /proc/net/bonding/bond0 and post the output, interested to see if the heartbeat is making it on to your nics. Tim This is my switch's configuration, it seems to be purely port-based multi-link. I can't post the output because as soon as i click Apply it just goes down and I can't navigate the webgui. To bring it back up I had to wipe the USB drive and reinstall everything. It sounds like your switch and the unRAID system don't understand each other, and hence a layer 2 loop is created. Start with connecting just ONE port and then apply the setting. You may need a console connection to verify the state of the ethernet port on the unRAID system. I'll try that since my MoBo has a serial port and I'll let you know
August 2, 20169 yr Author Your switch does not support dynamic LACP, it is another broken dell implementation. Id recommend the Cisco SG-200 series if you want a mondernish switch that supports most things. Ive upgraded to D-Link 10GB in the house because LACP is rather crappy. http://en.community.dell.com/support-forums/network-switches/f/866/p/19169209/19297716 Static LAG is your only option, if you loose one link your hosed. Tim Take some screenshots of the switch lacp config, there are different hashs for loadbalancing with multi-links, mac, mac + ip, ip+port, etc etc. Im interested to see what your min/max links are for this new lag, im assuming you set min/max at 4. When you switch the config on, can you look at your servers output of /proc/net/bonding/bond0 and post the output, interested to see if the heartbeat is making it on to your nics. Tim This is my switch's configuration, it seems to be purely port-based multi-link. I can't post the output because as soon as i click Apply it just goes down and I can't navigate the webgui. To bring it back up I had to wipe the USB drive and reinstall everything. Yeah... after a while I'm starting to realize that that think is a piece of junk... It claims a lot but in reality it sucks.
August 2, 20169 yr Yep, used to own a dell powerconnect, never again. Tim Your switch does not support dynamic LACP, it is another broken dell implementation. Id recommend the Cisco SG-200 series if you want a mondernish switch that supports most things. Ive upgraded to D-Link 10GB in the house because LACP is rather crappy. http://en.community.dell.com/support-forums/network-switches/f/866/p/19169209/19297716 Static LAG is your only option, if you loose one link your hosed. Tim Take some screenshots of the switch lacp config, there are different hashs for loadbalancing with multi-links, mac, mac + ip, ip+port, etc etc. Im interested to see what your min/max links are for this new lag, im assuming you set min/max at 4. When you switch the config on, can you look at your servers output of /proc/net/bonding/bond0 and post the output, interested to see if the heartbeat is making it on to your nics. Tim This is my switch's configuration, it seems to be purely port-based multi-link. I can't post the output because as soon as i click Apply it just goes down and I can't navigate the webgui. To bring it back up I had to wipe the USB drive and reinstall everything. Yeah... after a while I'm starting to realize that that think is a piece of junk... It claims a lot but in reality it sucks.
Archived
This topic is now archived and is closed to further replies.