Forums

Home / Forums

You need to log in to create posts and topics. Login · Register

Active/Backup bond not working


RafS
32 Posts
May 10, 2018, 9:34 pm

Hi,

I created a setup with with bond type 1 for my backend network.  Problem is that when the mail adaptor fail's the second one doesn't seem to take over.  when looking at the kern.logs it seems it doesn't catch the link failure.

Regards

Raf Schandevyl

#1

admin
2,930 Posts
May 13, 2018, 3:55 pm

 

We just tested it and it is working here.

can you check the mode is set to active-backup

cat /sys/class/net/BOND_NAME/bonding/mode

Are you using nics of different types, sometimes this causes issues in bonds.

Last edited on May 13, 2018, 3:56 pm by admin · #2

RafS
32 Posts
May 14, 2018, 5:46 am

Hi i'm using 2 identical 10G nics.  (Intel x552)

Output :

cat: /sys/class/net/BOND_NAME/bonding/mode: No such file or directory
root@ps-node-01:~# cat /sys/class/net/bBACKEND/bonding/mode
active-backup 1

Regards

Raf

#3

admin
2,930 Posts
May 14, 2018, 11:07 am

Can you check the backend networks were mapped to the bond not the nics. also if you can set the bond manually does it work ?

 

#4

RafS
32 Posts
May 14, 2018, 2:09 pm

How can i check if the backend networks is mapped to the bond and not the nic.  I the configuration interface i selected the bond.

I have not tested with creating the bond manually.

Regards

Raf

#5

admin
2,930 Posts
May 15, 2018, 3:42 pm

you can check your ips are mapped to bonds via

ip addr

can you also run the following to see the nics assigned to bonds:

cat /sys/class/net/BOND_NAME/bonding/slaves
cat /proc/net/bonding/BOND_NAME

The later should show you which nic is active in an active-backup bond

Last edited on May 15, 2018, 4:03 pm by admin · #6

RafS
32 Posts
May 15, 2018, 4:42 pm

when reading the output the ip addresses are on the bond. Everything seems ok, but when i disconnect the cable from the active slave, the otherone does not take over.  (Node goes down)

Regards

Raf

root@ps-node-01:~# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc mq master bMANAGEMENT state UP group default qlen 1000
link/ether ac:1f:6b:67:01:18 brd ff:ff:ff:ff:ff:ff
3: eth1: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc mq master bMANAGEMENT state UP group default qlen 1000
link/ether ac:1f:6b:67:01:18 brd ff:ff:ff:ff:ff:ff
4: eth2: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 9000 qdisc mq master bISCSI-01 state UP group default qlen 1000
link/ether ac:1f:6b:67:01:1a brd ff:ff:ff:ff:ff:ff
5: eth3: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 9000 qdisc mq master bISCSI-01 state UP group default qlen 1000
link/ether ac:1f:6b:67:01:1b brd ff:ff:ff:ff:ff:ff
6: eth4: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 9000 qdisc mq master bISCSI-02 state UP group default qlen 1000
link/ether ac:1f:6b:67:01:1c brd ff:ff:ff:ff:ff:ff
7: eth5: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 9000 qdisc mq master bISCSI-02 state UP group default qlen 1000
link/ether ac:1f:6b:67:01:1d brd ff:ff:ff:ff:ff:ff
8: eth6: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 9000 qdisc mq master bBACKEND state UP group default qlen 1000
link/ether ac:1f:6b:67:06:5e brd ff:ff:ff:ff:ff:ff
9: eth7: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 9000 qdisc mq master bBACKEND state UP group default qlen 1000
link/ether ac:1f:6b:67:06:5e brd ff:ff:ff:ff:ff:ff
11: bMANAGEMENT: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
link/ether ac:1f:6b:67:01:18 brd ff:ff:ff:ff:ff:ff
inet 192.168.100.150/24 scope global bMANAGEMENT
valid_lft forever preferred_lft forever
inet6 fe80::ae1f:6bff:fe67:118/64 scope link
valid_lft forever preferred_lft forever
12: bISCSI-01: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP group default qlen 1000
link/ether ac:1f:6b:67:01:1a brd ff:ff:ff:ff:ff:ff
inet6 fe80::ae1f:6bff:fe67:11a/64 scope link
valid_lft forever preferred_lft forever
13: bISCSI-02: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP group default qlen 1000
link/ether ac:1f:6b:67:01:1c brd ff:ff:ff:ff:ff:ff
inet6 fe80::ae1f:6bff:fe67:11c/64 scope link
valid_lft forever preferred_lft forever
14: bBACKEND: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 9000 qdisc noqueue state UP group default qlen 1000
link/ether ac:1f:6b:67:06:5e brd ff:ff:ff:ff:ff:ff
inet 192.168.160.150/24 scope global bBACKEND
valid_lft forever preferred_lft forever
inet 192.168.161.150/24 scope global bBACKEND
valid_lft forever preferred_lft forever
inet6 fe80::ae1f:6bff:fe67:65e/64 scope link
valid_lft forever preferred_lft forever
root@ps-node-01:~#
root@ps-node-01:~# cat /sys/class/net/bBACKEND/bonding/slaves
eth6 eth7
root@ps-node-01:~# cat /proc/net/bonding/bBACKEND
Ethernet Channel Bonding Driver: v3.7.1 (April 27, 2011)

Bonding Mode: fault-tolerance (active-backup)
Primary Slave: eth6 (primary_reselect always)
Currently Active Slave: eth6
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth6
MII Status: up
Speed: 10000 Mbps
Duplex: full
Link Failure Count: 2
Permanent HW addr: ac:1f:6b:67:06:5e
Slave queue ID: 0

Slave Interface: eth7
MII Status: up
Speed: 10000 Mbps
Duplex: full
Link Failure Count: 3
Permanent HW addr: ac:1f:6b:67:06:5f
Slave queue ID: 0
root@ps-node-01:~#

#7

admin
2,930 Posts
May 15, 2018, 10:01 pm

Turn off fencing from the maintenance page so the node does not get killed, this will make it easier to debug. Then if you unplug the eth6 cable, use the prev commands to see if eth7 gets active or not. I would recommend you also try to manually set the bonding and see if it is different. i would also check your switches and see if there is a settings that need to be done.

#8