10

I have set up two interfaces as bonding slaves (eth0 and eth1 bonded as bond0). How are people monitoring the status of the slaves?

How are you notified if the link fails on one interface? Do you poll something in /sys/class/net/bond0 or /proc/net/bonding/bond0 ? Is there a daemon I can configure to get notifications?

Some context: I'm running SUSE 11 server SP1 and I'm not interested in any graphical tool as this server is not running X11.

Louis Munro
  • 195
  • 1
  • 1
  • 7

4 Answers4

4

There are some ways to do this:

  1. As you mentioned, query status from /proc/net/bonding/bond0:

    if [ `grep -c down /proc/net/bonding/bond0` -eq 1 ]; then
    echo "`date +%c`" | \
    mail -s "$(grep -B1 down /proc/net/bonding/bond0 | head -1 | \
    awk -F': ' '{ print $2 }') is down" your@email.address 
fi

  2. Nagios's check_linux_bonding plugin.

  3. SNMP.

quanta
  • 50,327
  • 19
  • 152
  • 213
  • This will work only if 1 slave is down but you might want to receive e-mails if all of 2 slaves or 2 of 3 or more slaves are down. I suggest `if fgrep -q down /proc/net/bonding/bond0` to check for any occurrence of `down`. – Daniel Böhmer Nov 07 '17 at 15:56
  • In the end I put `fgrep -C999 down /proc/net/bonding/bond0 || true` in my crontab which will simply send the whole file content if `down` was found. – Daniel Böhmer Nov 07 '17 at 16:04
2

If you are doing bonding, check /proc/net/bonding/bond0.

You should see something like this 

Ethernet Channel Bonding Driver: v3.7.0 (June 2, 2010)

Bonding Mode: fault-tolerance (active-backup)
Primary Slave: None
Currently Active Slave: eth0
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth0
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:0d:60:d4:a3:00
Slave queue ID: 0

Slave Interface: eth1
MII Status: up
Speed: 100 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:0d:60:d4:a3:01
Slave queue ID: 0

I just check to see if the Currently Active slave has changed or not.

Also if you are just looking to see if link is lost on one of your NIC's, you can still do that though SNMP or other standard forms of monitoring.

Squidly
  • 1,685
  • 1
  • 14
  • 18
1

Under Nagios I used check_ifoperstatus on each of the slave interfaces which performs an SNMP query to 1.3.6.1.2.1.2.2.1.8 in the IF-MIB.

Mark Wagner
  • 17,764
  • 2
  • 30
  • 47
0

I haven't any Suse here to test the command. In fedora/redhat I use "ethtool slave_device_name" http://en.wikipedia.org/wiki/Ethtool and i see "Link detected:" parameter.

Giovanni
  • 1
  • 1