Networker

[Networker] About i500 failover and SCSIDEV in Legato

2012-07-24 05:29:24
Subject: [Networker] About i500 failover and SCSIDEV in Legato
From: Manel Rodero Blanquez <manel AT FIB.UPC DOT EDU>
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Date: Tue, 24 Jul 2012 11:28:51 +0200
Hello,

I'm trying to setup our Legato NetWorker setup (7.6.3.870 on Windows Server 2008 SP2) to be fault tolerant when a fail-over occurs with our library.

We have the following library and server:

- Quantum Scalar i500 with 2 x HP LTO-4 drives and an I/O Blade
- PowerEdge R610 with 2 x HBA QLE2460
- Switches Brocade 300

with the following setup:

- Zone Path A = i500Port1 + switch3port9 + switch3port8 + ServerHBA0
- Zone Path B = i500Port2 + switch4port9 + switch4port8 + ServerHBA1

as seen in this schema:

Scalar i500 +-- I/O Blade +-- Port 1 <-----------+
            |             |                      |
            |             +-- Port 2 <--------+  |
            |             |                      |
            |             +-- Port 3 <-----+  |  |
            |             |                |  |  |
            |             +-- Port 4       |  |  |
            |             |                |  |  |
            |             +-- Port 5 <--+  |  |  |
            |             |             |  |  |  |
            |             +-- Port 6    |  |  |  |
            |                           |  |  |  |
            +-- Drive 2 <---------------+  |  |  |
            |                              |  |  |
            +-- Drive 1 <------------------+  |  |
                                              |  |
                                              |  |
Switch 4                                      |  |
Brocade 300 +-- Port X                        |  |
            |                                 |  |
            +-- ...                           |  |
            |                                 |  |
            +-- Port 9 <----------------------+  |
            |                                    |
            +-- Port 8 <----------------------+  |
            |                                 |  |
            +-- ...                           |  |
            |                                 |  |
            +-- Port 1                        |  |
                                              |  |
                                              |  |
Backup Server +-- HBA QLE2460 #1 <------------+  |
              +                                  |
              +-- HBA QLE2460 #0 <------------+  |
                                              |  |
                                              |  |
Switch 3                                      |  |
Brocade 300 +-- Port 1                        |  |
            |                                 |  |
            +-- ...                           |  |
            |                                 |  |
            +-- Port 8 <----------------------+  |
            |                                    |
            +-- Port 9 <-------------------------+
            |
            +-- ...
            |
            +-- Port X

The I/O Blade can do a fail-over if, for example, it detects a problem between one of its ports and the switches. If, for example, the port 2 is down, the drives and the changers are presented to the backup server through the path using port 1.

The problem here is that the SCSIDEV that OS/Legato see are different for every path. For example, we can see the changer as 8.0.5 if the active port in the I/O Blade is the port 2. But I can see it as 4.0.5 if the active port is the 1. And more strange, I've seen the changer as 4.1.5 when the active port is the 1 in the I/O Blade.

As far we know, in the x.y.z SCSIDEV number, the "z" is the LUN of the element (and always is 0 for the I/O Blade port, 1 for the first drive, 2 for the 2nd drive and 5 for the changer).

We would like to have a fixed x.y so the library definition in Legato doesn't need to be changed if a failure occurs (well, we haven't seen failures in production but we would like to automate a such situation).

Possible solutions:

a) Configure the different components of this setup so Windows/Legato always see the same SCSIDEV

b) Have an script that looks for the current SCSIDEV por the "Autochanger" (a simple 'inquire' will do that) and execute some command to disable the library, change the number and enable it again.

c) Other??

Do you have a similar setup? Is it possible to solve this "problem"?

Thanks for your help.

--

       Manel Rodero Blánquez
o o o  IT Systems Manager
o o o  inLab FIB
o o o  Barcelona School of Informatics
U P C  Universitat Politècnica de Catalunya - Barcelona Tech

       E-mail : manel AT fib.upc DOT edu
       Tel.   : +34 93 401 0847
       Web    : http://inlab.fib.upc.edu/

======================================================================

Abans  d'imprimir aquest missatge, si us plau, assegureu-vos que sigui
necessari. El medi ambient és cosa de tots.

--[ http://www.fib.upc.edu/disclaimer/ ]------------------------------

ADVERTIMENT  /  TEXT  LEGAL:  Aquest  missatge pot contenir informació
confidencial  o  legalment protegida i està exclusivament adreçat a la
persona  o entitat destinatària. Si vosté no es el destinatari final o
persona  encarregada  de  recollir-lo, no està autoritzat a llegir-lo,
retenir-lo, modificar-lo, distribuir-lo, copiar-lo ni a revelar el seu
contingut.  Si ha rebut aquest correu electrònic per error, li preguem
que  informi  al  remitent  i elimini del seu sistema el missatge i el
material annex que pugui contenir. Gràcies per la seva col·laboració.

<Prev in Thread] Current Thread [Next in Thread>
  • [Networker] About i500 failover and SCSIDEV in Legato, Manel Rodero Blanquez <=