strange I/O errors with SAN storage

[christian.peper at kpn.com]

Sijis,

thanks for the details! :)
We managed to get it fixed, but I still don't know what caused it.
We replaced the FC wires, switch ports and LUNs on the SAN. we suspected
the cabling or switch port configuration. But the switch worked.

I have never had to reinstall PP after a kernel upgrade. We are using PP
5.0.0 BTW.
EMCpower.LINUX-5.0.0-157.rhel.x86_64

dmesg of /var/log/messages was of no use because we never got there. It
got stuck during booting and enumerating all the devices (7 LUNs). The
machine was a fresh install, trying to do its first boot after
installation. It never made it. I think it was the FC cabling; some dirt
or bend or scratch or so.

And yes, I had the same HBA BIOS setttings. :)
Thanks again,
Chris.

> Date: Thu, 31 Jul 2008 14:08:19 -0500
> From: "Sijis Aviles" <saviles at empire-today.com>
> Subject: RE: strange I/O errors with SAN storage
> To: <linux-poweredge at dell.com>
> Message-ID:
> <D8D21A05D071AD4EBC94315FDD58A80A0188373C at NLILEXMB01.empirecarpet.com>
> Content-Type: text/plain; charset="us-ascii"
> 
> Robert,
> 
> Did you have have to reinstall PowerPath after a kernel 
> upgrade? We are using PP 4.5.1 and everytime we upgrade the 
> kernel, I have to reinstall the software.
> 
> Chris,
> 
> I've seen similar errors on my RHEL4 servers when the path 
> dies because the path to the SAN dies (HBA, cable, switch 
> port) . From looking at your errors, it seems like it could 
> be connectivity issue. Try changing the FC cable, it might be 
> bad or seeing if there are any errors in the switch.
> Does dmesg or /var/log/messages show anything useful?
>  
> We are using PE2950's and we use local disk for OS and SAN 
> for storage. 
> 
> As a note, in our documentation, this is our config for the 
> HBA to connect to our EMC Symmetrix.
> a. Host Adapter BIOS = Enabled
> b. Connection Options = 1 (Point to point only) c. Data Rate 
> = 1 (2GB/S) Everyone else is defaults