SSA_LINK_OPEN

[srahma]

Hi,
We have RS/6000 S7A with AIX 4.3.3, clustered using HACMP, having 2 SSA storage linked via Fiber optic.
We are constantly getting <SSA_LINK_OPEN Error> and it is causing some disks to go into STALE or Missing status.
What might be causing this? Any solution? anyone had similar problem?..
We have contacted IBM, they came and played with connections and cables, replaced fiber optic and so on....
still same if not worst.

Thanks for your feedback,
Salam

 

[PSD]

Salam,

Have you had a look through diags - SSA services aids - configuration verfication / loop verification.

I have had a number of link errors in the past - are you getting them at the same time each day? or is there another pattern. The reason is that the cron for root has SSA diags. and if the time is sync'd on the HA boxes then the diags. are run at the same time on the shared SSA loops and errors can occur.

It may be that a cable or fiber optic link is on the way out, how long are the cables? In the past I have has this when a) you reboot a server b) a cable is failing....

Another thing to look at is what level of microcode are you on for adapters and disks i.e.:

ssadload -s - this shows the disks, you will get errors in HACMP config. as some disks may be varyed onto the other node.

lsattr -El ssa0 - ROS level and ID.

Just a starting point. PSD
IBM Certified Specialist - AIX V4.3 Systems Support
IBM Certified Specialist - AIX V4 HACMP

 

[Jaganr]

Hello Salam,

We also faced this error for quite some time.

Following were done:

a. FO cables when tested were found giving heavy db loss
around 9.5 when the acceptable value is 3
b. Introduced the fibre optic conditioners
c. Cleaned the FO adapter connectors

After doing the step c. we did not get any error in the last four days. Previously we used to get ssa link open errors almost every day several times.

Jagan