I have a 10-switch Brocade Silkworm 2800 with firmware version 2.3. I would like to know what would happen if we need to upgrade or downgrade the firmware version in four switches. Do you think this would be a big problem with the ther switches? We need to change the firmware version because we have problems with the storage (ESA12000, (2)EMA8000). In other switches, we have connected two EMCs. All switches are interconnected (ISL).
Brocade recommends using the same firmware versions on switches that are connected via E-ports (ISL'ed)to create a single fabric. You should plan for some downtime to do this since the hosts connected to the switches that are updated will experience a loss in session as you reboot the switch. (You need to reboot the switch for the new firmware to take effect.)
As always, call your vendor for information on what is supported before you proceed or let them help you do it.
Before you upgrade the firmware, use the "configUpload" command to save your zone information to a server on your network. Use a separate filename for each switches configuration. The data is saved as a text file that can be edited. Use the switchname or switch DomainID as the name of the config file so you can reference it to the switch.
The command to download firmware to a Brocade switch is as follows:
>firmwareDownload ["host","user","file" [,"passwd"]]
You must have "admin" rights on the switch to do this. The latest version can be downloaded from the support pages of the Brocade Web site (for the 2800, that's version 2.6)
This is from the manual:
"The download process uses either FTP (File Transfer Protocol) or the RSHD protocol (TCP service 514). To use RSHD on Windows NT or 9x, two utilities are supplied with the firmware file, RSHD.EXE and CAT.EXE, together with instructions on how to install and run them. The FTP server or RSHD must be running before a firmware download can occur."
After a download successfully completes, the switch must be rebooted to activate the new firmware.
Here is an example:
55696+6984+133172, csum 7eca
writing flash 0 ..................
writing flash 1 ..................
I would suggest in the future you redesign your fabric by splitting it in half thereby creating two separate fabrics. You should then use TWO HBAs in every server and connect one HBA to each fabric. This will let you load balance and provide failover capability to your storage. This way when you need to service a component, like your switches, or Storageworks controllers, you will not impact production. All data will automatically flow across the unaffected path.
The good news is that all Brocade firmware is usually backward compatible so you should be able to do the upgrade on the Storageworks switches without having problems with the EMC switches.
Editor's note: Do you agree with this expert's response? If you have more to share, post it in one of our .bphAaR2qhqA^0@/searchstorage>discussion forums.
This was first published in May 2002