FW_120_StudentGuide_l
FW_120_StudentGuide_l
FW_120_StudentGuide_l
Revision 0118 1 - 10
FW-120 Firmware Management
Footnote 1: If this switch was purchased from an OEM vendor there may be upgrade
requirements different from those listed here. Contact your vendor for additional information.
Footnote 2: In other words, upgrading a switch from Fabric OS v7.3.0 to v8.0.1 is a two-
step process—first upgrade to v7.4.0, and then upgrade to v8.0.1. If you are running a pre-
Fabric OS v7.3.0 version you must upgrade to v7.3.0, then to v7.4.0, and finally to v8.0.1.
Footnote 3: Ensure that all serial consoles (both CPs for directors) and any open network
connection sessions, such as Telnet, are logged and included with any trouble reports.
The firmware download process is the same for all of Brocade’s SAN switches. Firmware is
stored in flash on two separate partitions, primary and secondary. Firmware is first
downloaded to the secondary partition and the switch rebooted from the updated code. This
allows an opportunity to asses the new firmware and ensure that there are no problems.
Once the switch has booted successfully from the update code it is then copied to the
remaining flash partition.
The firmware download process for a single-CP switch is less complex than the Brocade
directors.
The flash is divided into two equal size partitions. The version of firmware running on the
switch is in the Primary Partition. The version in the Secondary can be the same as the
Primary or different.
The following slides will illustrate the individual steps that occur when a
firmwaredownload command is issued.
Upon completion of the download, the Secondary Partition becomes the Primary and the
Primary Partition becomes the Secondary.
After rebooting, the new version is now running on the switch. The reboot performed in this
case is an hareboot which is non-disruptive to SAN traffic.
Revision 0118 2 - 10
FW-120 Firmware Management
Revision 0118 2 - 11
FW-120 Firmware Management
SW1:admin> firmwareshow
Appl Primary/Secondary Versions
------------------------------------------
FOS v8.0.1
v8.0.1
Revision 0118 2 - 12
FW-120 Firmware Management
Revision 0118 2 - 13
FW-120 Firmware Management
Footnote 1: If the hashow command output shows any problems with the HA status of the
CPs do not continue with any firmware upgrades until the issue is resolved.
The hashow command displays many details about the current high-availability status of the
Brocade director including:
• Local CP state (slot number, CP ID): Active or Standby
• Remote CP state (slot number, CP ID): Active or Standby
• High Availability: HA Enabled or HA Disabled
• Heartbeat: Up or Down
• Health of standby CP: Healthy (standby CP is running, and the health monitor has not
detected any errors), Failed (standby CP is running, but the health monitor has
discovered a problem with the blade), and Unknown (the standby CP does not exist,
the heartbeat is down, or the health monitor detects a configuration file error).
• HA synchronization status: HA State Synchronized (the system is currently fully
synchronized; a CP Card failover would be non-disruptive) and HA State Not In Sync
(the system is unable to synchronize the two CPs, due to the standby CP being faulty
or another system error; a CP Card failover would be disruptive).
Revision 0118 2 - 14
FW-120 Firmware Management
Active CP state:
• Cold Recovered — The CP became the Active CP through a power on or hard reboot
and is disruptive.
• Warm Recovered — The CP became the Active CP through a CP failover and is non-
disruptive.
Standby CP state:
• Healthy — The Standby CP is running and the background health diagnostic has not
detected any errors.
• Failed — The Standby CP is running but the background health diagnostic has
discovered a problem with the blade. Failover is disabled until the Standby CP is
repaired. Information about the failure for the Standby CP is displayed.
• Unknown — The Standby CP health state is unknown for one of the following reasons:
Standby CP does not exist, heartbeat is down
The Health Monitor has detected a configuration file error
Revision 0118 2 - 15
FW-120 Firmware Management
Footnote 1: The hareboot command can also be used and provides the same
functionality as the hafailover command (Force a manual CP Card failover from the
Active CP.
Revision 0118 2 - 16
FW-120 Firmware Management
Revision 0118 2 - 17
FW-120 Firmware Management
Revision 0118 2 - 18
FW-120 Firmware Management
Active CP state:
• Cold Recovered — The CP became the Active CP through a power on or hard reboot
and is disruptive.
• Warm Recovered — The CP became the Active CP through a CP failover and is non-
disruptive.
Standby CP state:
• Healthy — The Standby CP is running and the background health diagnostic has not
detected any errors.
• Failed — The Standby CP is running but the background health diagnostic has
discovered a problem with the blade. Failover is disabled until the Standby CP is
repaired. Information about the failure for the Standby CP is displayed.
• Unknown — The Standby CP health state is unknown for one of the following reasons:
Standby CP does not exist, heartbeat is down
The Health Monitor has detected a configuration file error
Revision 0118 2 - 19
FW-120 Firmware Management
Revision 0118 2 - 20
FW-120 Firmware Management
The firmware download process for a Brocade director is more complex than other Brocade
switches. The director has dual Control Processor cards (CP0 and CP1) with one
performing an active role and the other a standby role.
Each CP card has two compact flash cards that hold the Fabric OS firmware. The version of
the firmware controlling the switch is in the primary partition and another version of firmware
is kept in the secondary partition. The version in the secondary partition can be the same as
the primary or different.
This is what makes the firmware process more complex. The following slides will illustrate
the individual steps that occur when a firmwaredownload command is issued on a
Brocade director.
Revision 0118 2 - 21
FW-120 Firmware Management
Revision 0118 2 - 22
FW-120 Firmware Management
The Active CP instructs the Standby CP to perform a firmware download of version Fabric
OS B. The firmware is downloaded into the secondary partition. If the Standby CP does not
respond within 10 minutes, the Active CP will time out and abort the command.
All Brocade directors download firmware to the Standby CP card first. Before the firmware
download process begins, Fabric OS checks to see that both CP cards are running the
same version of Fabric OS. If they are running different versions, each CP card will have to
be updated separately.
Revision 0118 2 - 23
FW-120 Firmware Management
At this point the new firmware has been copied to the secondary partition on the Standby
CP. The CP prepares to reboot with the new code by swapping the pointers for the primary
and secondary partitions. The secondary partition becomes the primary and vice-versa.
Revision 0118 2 - 24
FW-120 Firmware Management
Before the active CP can instruct the standby CP to reboot, the standby CP has to tell the
active CP that the checksum is good.
After the pointers are swapped, the active CP instructs the standby CP to reboot. The active
CP remains operational and the chassis is still running that code. When the standby CP
reboots, it loads the new firmware from the primary partition of the compact flash card. The
standby CP is now running the upgraded Fabric OS code.
Revision 0118 2 - 25
FW-120 Firmware Management
The standby CP is now running the upgraded Fabric OS code. It sends status back to the
Active CP that the reboot was successful. The CP heartbeat is up and the CPs are
communicating with each other.
Note: The CP verifies the hasync and heartbeat is up before an hafailover occurs.
Revision 0118 2 - 26
FW-120 Firmware Management
With the Standby CP now running the new firmware, a failover is performed. The Standby
CP becomes Active and the Active CP becomes the Standby.
Revision 0118 2 - 27
FW-120 Firmware Management
The Active CP copies the firmware from the primary partition to the secondary partition of
the Standby CP using the internal network between the CP blades.
Revision 0118 2 - 28
FW-120 Firmware Management
Revision 0118 2 - 29
FW-120 Firmware Management
Before the active CP can instruct the standby CP to reboot, the standby CP has to tell the
active CP that the checksum is good.
The Standby CP then reboots in order to load the new Fabric OS firmware. At this point both
CPs are running the new firmware but it has only been copied to one compact flash. This
ensures a level of redundancy, if the upgrade process fails for some reason the switch can
revert to the old firmware.
Revision 0118 2 - 30
FW-120 Firmware Management
Once both CPs are running the new firmware the Active and Standby CPs perform a
firmwarecommit. This copies the firmware from the primary partition to the secondary
compact flash partition.
Revision 0118 2 - 31
FW-120 Firmware Management
Once the firmwarecommit is completed the upgrade process is completed. The new
firmware is stored on the primary and secondary partitions of both CP blades, and both CPs
are now running the new firmware.
Notice that the CPs have switched roles at the end of the upgrade process. It is possible to
revert back to the original Active/Standby roles by issuing the hashow command to verify
the CPs are in sync and then the hafailover command.
Revision 0118 2 - 32
FW-120 Firmware Management
Revision 0118 2 - 33
FW-120 Firmware Management
The firmwaredownload status command can be used to monitor the progress of the
firmware upgrade at any time during the process.
DCX8510:admin> firmwaredownloadstatus
[1] Fri May 23 17:43:00 2016
Slot 6 <CP0, active>: Firmware is being downloaded to standby CP.
This step may take up to 30 minutes.
[2] Fri May 23 17:51:02 2016
Slot 6 <CP0, active>: Firmware has been downloaded successfully to
Standby CP.
[3] Fri May 23 17:51:06 2016
Slot 6 <CP0, active>: Standby CP is going to reboot with new
firmware.
[4] Fri May 23 17:53:08 2016
Slot 6 <CP0, active>: Standby CP booted successfully with new
firmware.
[5] Fri May 23 17:54:26 2016
Slot 7 <CP1, active>: Forced failover succeeded. New Active CP is
running new firmware
[6] Fri May 2 3 17:55:00 2016
Slot 7 <CP1, active>: Firmware is being downloaded to standby CP.
This step may take up to 30 minutes.
[7] Fri May 23 18:00:50 2016
Slot 7 <CP1, active>: Firmware has been downloaded successfully on
Standby CP.
[8] Fri May 23 18:00:54 2016
Slot 7 <CP1, active>: Standby CP reboots.
[9] Fri May 23 18:02:35 2016
Slot 7 <CP1, active>: Standby CP rebooted successfully with new
firmware.
[10] Fri May 23 18:02:37 2016
Slot 7 <CP1, active>: Firmware commit operation has started on both
active and standby CPs.
[11] Fri May 23 18:02:37 2016
Slot 7 <CP1, active>: The firmware commit operation has started.
This may take up to 10 minutes.
[12] Fri May 23 18:07:21 2016
Slot 7 <CP1, active>: The commit operation has completed
successfully.
[13] Fri May 23 18:07:21 2016
Slot 7 <CP1, active>: Firmware commit operation has completed
successfully on active CP.
[14] Fri May 23 18:07:21 2016
Slot 7 <CP1, active>: Firmwaredownload command has completed
successfully. Use firmwareshow to verify the firmware versions.
Revision 0118 2 - 34
FW-120 Firmware Management
Revision 0118 2 - 35
FW-120 Firmware Management
Note: You can also import the release notes, in PDF format, at this time.
Revision 0118 2 - 36
FW-120 Firmware Management
Once the firmware has been imported you can select it to view detailed information such as
the release date and the import date. Additionally Network Advisor can show you which
switches are capable of, or are running, the imported firmware. Two radio buttons are
available for displaying switch and firmware information:
Revision 0118 2 - 37
FW-120 Firmware Management
2
1
Revision 0118 2 - 38
FW-120 Firmware Management
Revision 0118 2 - 39
FW-120 Firmware Management
3
5
Revision 0118 2 - 40
FW-120 Firmware Management
Updating switch firmware is done from the Download tab. Select the switches you wish to
update and click the right arrow to add them to the Selected Switches list. You can use
SHIFT+click to select a range of devices or CTRL+click to select multiple devices
individually.
Revision 0118 2 - 41
FW-120 Firmware Management
Once the switches have been selected you can select the firmware version you wish to
update to from the Firmware to Download drop-down next to each switch. Only firmware
that has been imported into the repository and is supported by that product will be available
for selection. Alternatively you can use the Select Latest button to select the latest
applicable firmware version for each switch.
By default Network Advisor performs firmware updates in a parallel fashion, by sending the
firmware to each switch simultaneously. If you would prefer that the update be done serially,
one switch at a time, check the Serial download box at the top of the window.
Also, by default, Network Advisor will not downgrade a switch or flash the same version of
firmware to a switch. This behavior can be changed by checking the Overwrite current
firmware box. This tells Network Advisor to do the firmware upgrade even on switches that
have the same or newer revisions.
To begin the firmware download process click the Download button. You will receive a
warning not to power cycle the switches during the firmware download process.
Revision 0118 2 - 42
FW-120 Firmware Management
Footnote 1: The graphic below indicates the change in progress from the graphic on the
slide above. The messages displayed in the Status Details window are the same messages
displayed when using the firmwaredownloadstatus command from the CLI.
Revision 0118 2 - 43
FW-120 Firmware Management
Revision 0118 2 - 44
FW-120 Firmware Management
Revision 0118 2 - 45
FW-120 Firmware Management
Only when the –s argument is added to the firmwaredownload command are the
following two options available:
• Autocommit mode. If the –n operand is added to the command, then after the new
firmware becomes the Primary Partition, it is not copied to the Secondary Partition.
The default is to commit the firmware.
• Autoreboot mode. If the –b operand is added to the command, then an HAreboot is
performed after the firmware download completes. The default is to not reboot the
firmware.
• Notice that the example above was recorded on a dual-CP Brocade DCX8510. The only
operand included at the command line is –s, so users are prompted to provide the FTP
server IP address, user name, firmware filename, and password, as well as the two single-
CP options.
Revision 0118 2 - 46
FW-120 Firmware Management
Revision 0118 2 - 47
FW-120 Firmware Management
Attempting to mount an unsupported USB device will result in the following failure:
DCX8510:admin> usbstorage -e
Fail to enable USB storage device. Error:Device is not found
Revision 0118 2 - 48
FW-120 Firmware Management
Revision 0118 2 - 49
FW-120 Firmware Management
The auto-leveling process automatically ensures compatibility between the version of Fabric
OS running on the Active CP and all of the supported AP blades installed in the director
chassis. When the Brocade DCX family product boots up or a supported AP blade is
inserted or powered on, the Active CP determines if the installed AP blades are running the
same version of Fabric OS as the active CP. If the Fabric OS versions differ, the Active CP
will install the Fabric OS image on all installed advanced blades. Auto-leveling occurs
automatically, without administrator intervention.
Auto-leveling occurs:
• After CP failover (CP error), including during a dual-CP firmware download (e.g.
firmwaredownload -s)
• After hot-plug of a supported AP blade or when slotpoweron is invoked on a slot
containing one of these blades
• The active CP always updates the AP blade.
• Auto-leveling is always enabled, and cannot be disabled
Revision 0118 2 - 50
FW-120 Firmware Management
Revision 0118 2 - 51
FW-120 Firmware Management
Revision 0118 2 - 52
FW-120 Firmware Management
Revision 0118 2 - 53
FW-120 Firmware Management
Revision 0118 2 - 54