APA failovers

[Edward Ady]

We are running port aggregation on several systems. Tested it out and it works great, only one problem. How do we tell when we fail only one port in a apa group?

[Steven E. Protter]

It should show up as a message in /var/adm/syslog/syslog.log

EMS might be configured to email on issues as well.

SEP

[Mel Burslan]

also dmesg output will show a failed lan interface entry in it.
if you have a monitoring solution watching over either of syslog.log or dmesg output periodically, it will catch it. If you don't you can write a small script and run it every few minutes to see if such a thing happened and if it did you can try to warn some people via email or something like that.

[Edward Ady]

We unplugged one of the ports and there was no message in syslog or on console. The only time we saw a message is when we unplugged both ports.

We did find something which can tell if we have a failover. If you have two ports in an apa group and they are both working, they do not show up in a lanscan, but it does not show up in syslog/dmesg.

Should a message be logged if one port is unplugged ?

Could this be an issue with the level of logging that is defined in /etc/syslog.conf

..Ed

[Xianjie Zhang · Accepted solution]

APA normally puts log in NetTL logs. (man nettl).
If a failover happens and there are still up ports, it is logged in nettl as a warning message,which is turned off by default.
To turn it on, run "nettl -log w -e hp_apa"
To view the logs, run "netfmt nettl.LOG000". (Note: the exact log filename may vary).
NetTL also has other nice features for filtering. Please check out the man page.
Below is the sample messages when I unplug port 4:

**********************Auto-Port Aggregation/9000 Networking*****************@#%
Timestamp : Thu Jun 24 MDT 2004 10:12:06.421375
Process ID : [ICS] Subsystem : HP_APA
User ID ( UID ) : -1 Log Class : WARNING
Device ID : 4 Path ID : 0
Connection ID : 0 Log Instance : 0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<3002> HP Auto-Port Aggregation product detected port 4 went down in
link aggregation 902.

**************************100 Mb/s LAN/9000 Networking**********************@#%
Timestamp : Thu Jun 24 MDT 2004 10:12:06.421432
Process ID : [ICS] Subsystem : BTLAN
User ID ( UID ) : -1 Log Class : ERROR
Device ID : 4 Path ID : 0
Connection ID : 0 Log Instance : 0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<7002> 10/100BASE-T driver detected bad cable connection between
the adapter in slot(Crd In#) 4 and the hub or switch.


**********************Auto-Port Aggregation/9000 Networking*****************@#%
Timestamp : Thu Jun 24 MDT 2004 10:12:09.647289
Process ID : [ICS] Subsystem : HP_APA
User ID ( UID ) : -1 Log Class : DISASTER
Device ID : 902 Path ID : 0
Connection ID : 0 Log Instance : 0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<1006> HP Auto-Port Aggregation product found that ports in failover
group lan902 are no longer connected to each other. Port 5 did
not receive any poll packets.

[Xianjie Zhang]

Even if you did not turn on "warning" log in nettl, you can still see "error" and "disaster" logs (which are turned on by default). The only log you will miss in the above example (see my previous post) without turning on "warning" log is:
**********************Auto-Port Aggregation/9000 Networking*****************@#%
Timestamp : Thu Jun 24 MDT 2004 10:12:06.421375
Process ID : [ICS] Subsystem : HP_APA
User ID ( UID ) : -1 Log Class : WARNING
Device ID : 4 Path ID : 0
Connection ID : 0 Log Instance : 0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
<3002> HP Auto-Port Aggregation product detected port 4 went down in
link aggregation 902.

[Edward Ady]

Thanks.

Just what I was looking for.