HBA's DIsconnect and Reconnect Messages

[Nate Arnold]

Hi. I was investigating a data integrity error and noticed the following messages in /var/adm/messages

1/0/2/1/0: Device at device id 0x21100 is back in Name Server GPN_FT (FCP type)
response, and its 'Port World-Wide Name' remains the same as
original.
device id = loop id, for private loop devices
device id = nport ID, for fabric/public-loop devices
System will be able to see LUNs behind this port
(might need to run 'ioscan' first).

0/0/10/1/0: Device at device id 0x10100 is back in Name Server GPN_FT (FCP type)
response, and its 'Port World-Wide Name' remains the same as
original.
device id = loop id, for private loop devices
device id = nport ID, for fabric/public-loop devices
System will be able to see LUNs behind this port
(might need to run 'ioscan' first).

These messages repeat themselves every ten minutes. We are hard port zoned and a review of the switches show no errors and only connections to our XP12000.

Any suggestions on how to stop logging these messages?

Thanks

[Steven E. Protter]

Shalom,

/var/adm/messages ? That a new file to me. /var/adm/syslog/syslog.log ?

Perhaps a new file for HP-UX 11.31 ?

OS and version?

Type of HBA?

More complete info leads to better solutions here.

You don't want to stop this logging.

You want to find the source of this problem and correct it.

You have a disconnect and a data integrity problem.

Check the SAN, disk arrry, cabling, and don't stop looking until you correct this problem.

This type of problem usually results in a call to HP hardware, you might want to log such a call.

SEP

[likid0 · Accepted solution]

And also have a look at the stats:

[root@sdp:~]# fcmsutil /dev/fcd0 stat
Tue Nov 25 08:25:11 2008
Channel Statistics

Link Statistics from firmware ...
Link Failure count 0 Loss of Sync count 1
Loss of Signal count 1 Prim seq proto err count 0
Invalid Word count 0 Invalid CRC count 0
SBC hit counter 0



look at the lost of Sync,Signal and the CRC count.


you can also look at the stats by nport:

[root@sdp:~]# fcmsutil /dev/fcd devstat all

[Nate Arnold]

Hi. The messages are being logged in /var/adm/messages. /var/adm/syslog/syslog.log is not logging any errors.

This is an 11.23 version of HP-UX.

The HBAs are HP AB379-60001 4Gb Dual Port PCI/PCI-X Fibre Channel Adapter

The integrity check of the database comes up clean so I think we are logging spurious errors but I would like to get to the bottom of this.

This servers is a member of a two node cluster. The second server in the cluster is experiencing the same level of messaging.

Review of fc statistics don't show any CRC errors, loss of Sync nor loss of signal.

[Florian Heigl (new acc)]

1. if you're having issues, then please don't stop logging error messages ;)
2. review your FC nameserver status on the switches, this is not a normal connectivity issue and not normal behaviour, unfortunately I don't know a lot about the gory bits of FC (or iSNS just alike) to point you at something.

Also have a look at the stats for ns_open in fcmsutil /dev/xxxx stat
Fabric Statistics ...
[...]
ns_open failures 0

Good luck,
florian