firmwarecommit/firmwaredownload failed during Brocade[TM] Silkworm fabric OS upgrade

Asset ID:	1-75-1012659.1
Update Date:	2013-03-15
Keywords:

Solution Type Troubleshooting Sure

Solution 1012659.1 : firmwarecommit/firmwaredownload failed during Brocade[TM] Silkworm fabric OS upgrade

Related Items


Brocade 24000 Director
 Brocade 12000 Fabric Switch
 Brocade 48000 Director

Related Categories


PLA-Support>Sun Systems>DISK>Switch>SN-DK: Brocade Switch
 _Old GCS Categories>Sun Microsystems>Switches>Brocade

PreviouslyPublishedAs
217411

Description

firmwaredownload failed on 12K,24K,48K switches

firmwaredownloadstatus indicates firmwarecommit failed.

Steps to Follow

This failure condition can occur on Brocade switches that utilise dual FW partitions such as the 12K, 24K and 48K switch.

Step 1: Run firmwareshow and firmwaredownloadstatus to identify the status of the firmware.

>> Depending on the options selected during firmwaredownload, the output may show a difference between the active and secondary partitions as per the following example. This will also occur if the firmwarecommit stage of the download process failed.

admin> firmwareshow
Local CP (Slot 5, CP0): Active
Primary partition: v5.0.4a
Secondary Partition: v4.4.0e
Remote CP (Slot 6, CP1): Standby
Primary partition: v5.0.4a
Secondary Partition: v5.0.4a

>> Should the issue exist, indication that firmwarecommit failed will be displayed similar to the following firmwaredownloadstatus output :-

admin>firmwaredownloadstatus
.....
[2]: Sat Jul 22 22:28:13 2006
cp1: Firmwarecommit has started.

[3]: Sat Jul 22 22:30:51 2006
cp1: Firmwarecommit failed.

>> Attempting to run firmwarecommit on the affected CP will result in failure, similar to the following:-

admin:firmwarecommit
.....
rsync: error writing 69 unbuffered bytes - exiting: Broken pipe
rsync error: error in rsync protocol data stream (code 12) at io.c(465)
Error: (Firmwarecommit failed.)

Step 2: This kind of error will occur if the secondary partition Compact Flash(CF) is full. Check to see if the affected CP has a full secondary partition

>> On the affected CP, run "supportshow" command and search for the 'df' output , similar to the following

/bin/df:
Filesystem 1k-blocks Used Available Use% Mounted on
/dev/root 120112 71152 48960 59% /
/dev/hda2 120128 117725 2402 98% /mnt

>> This indicates that the secondary partition ( /mnt) Compact Flash(CF) is almost full.

Note:If you have root password on the target switch, you can directly run the command "/bin/df"

  Care should be exercised when logged in as root user. Some commands may impact the switch's normal operations.

Step 3: The common cause of the partition filling up is through the storage of old core images. Run savecore command to check for the existence of core files.

admin> savecore -l
/mnt/core_files/nsd/core.837
/mnt/core_files/panic/core.837

Step 4: Using the saveCore command is the only way to remove core files; it makes more space available, in order to remove all core files in all directories, command "saveCore -R" can be used.

admin> savecore -R
removing core files under: /mnt/core_files/nsd
removing core files under: /mnt/core_files/panic

admin> savecore -l
No core files found!

Step 5: Check current secondary partition Compact Flash(CF) usage status as per step (2)

/bin/df:
Filesystem 1k-blocks Used Available Use% Mounted on
/dev/root 120112 71152 48960 59% /
/dev/hda2 120128 70316 49812 59% /mnt

Step 6: Run firmwarecommit on the affected CP and verify it's completion using firmwaredownloadstatus

admin> firmwarecommit
Validating primary partition...
Doing firmwarecommit now.
Please wait ...
Replicating kernel image.
................
Firmwarecommit completes successfully.

>> Note: firmwarecommit command can take a long time to complete.

admin> firmwaredownloadstatus
......

[4]: Sat Jul 22 23:35:27 2006
cp1: Firmwarecommit has started.

[5]: Sat Jul 22 23:45:35 2006
cp1: Firmwarecommit has completed successfully.

Step 7: Run firmwareshow to verify the FW on all CP partitions is as expected, as per the following example.

admin> firmwareshow
Local CP (Slot 5, CP0): Active
Primary partition: v5.0.4a
Secondary Partition: v5.0.4a
Remote CP (Slot 6, CP1): Standby
Primary partition: v5.0.4a
Secondary Partition: v5.0.4a

>> Note: The FW versions detailed in this document are examples only.

As per the Brocade document Firmware Download: Best Practice Notes, a check should be made prior to attempting firmwaredownload to ensure that the compact flash is lower than 90% full

Product
Brocade 12000 2 GB Switch
Brocade 48000 Director
Brocade SilkWorm 24000 Director

Internal Comments

The Brocade document can be obtained from http://pts-storage.us.oracle.com/products/Switch/local/brocade/fwupdate_bestpractice.pdf

Brocade, 12000, 24000, 48000, firmwaredownload, firmwarecommit, firmwaredownloadstatus
Previously Published As
86516

Change History
Date: 2006-09-25
User Name: 31620
Action: Approved
Comment: Verified Metadata - ok
Verified Keywords - ok
Verified still correct for audience - currently set to contract
Audience left at contract as per FvF at
http://kmo.central/howto/content/voyager-contributor-standards.html
Checked review date - currently set to 2007-08-02
Checked for TM - ok as it is
Publishing under the current publication rules of 18 Apr 2005:
Version: 5
Date: 2006-09-22
User Name: 31620
Action: Accept
Comment:
Version: 0
Date: 2006-09-22
User Name: 34660
Action: Approved
Comment: I performed a MAJOR rewrite. Although this should be an SRDB I'm not going to delete and recreate it.
Version: 0
Date: 2006-09-22
User Name: 125432
Action: Approved
Comment: Brian,

Modified this document as you suggested, Please review it again,

Thanks,
Willis
Version: 0
Date: 2006-09-19
User Name: 34660
Action: Rejected
Comment: I dont like the following statement.
"Based on Brocade[TM] support team feedback"

Given that you have set this to 'contract' it will give the impression to customers that Sun doesn't know how to support the product.
Version: 0
Date: 2006-09-19
User Name: 34660
Action: Accept
Comment:
Version: 0
Date: 2006-09-19
User Name: 34660
Action: Add Comment
Comment: I think this is an SRDB as it has a clear symptom and resolution.

The Problem statement is far too long. It should be reduced to include that fact that on the switch, firmwaredownload failed.

You probably want to list the switches where this could occur.
root account use is dangerous, it needs a warning and should not be used unless absolutely required.
you can get the df output from a supportshow/supportsave , so checking for the free space can be done there.
It would be good if you had an indication of roughly how much free space is required so the customer/field can check prior to hitting the condition.

I think you should also log an RFE. If the switch knows how big the image is and how much space it has then it can work it out itself - and provide a warning/error message with the cause.
Version: 0
Date: 2006-09-19
User Name: 125432
Action: Approved
Comment: KE experts,

Follow KE suggestion, send this TS doc to Storage group,
Please kindly review it.

Thanks & Rgds,
Willis
Version: 0
Date: 2006-09-18
User Name: 95826
Action: Rejected
Comment: Hi WIllis,

it seems that this document was sent to the wrong techgroup. 'Brocade' related articles should go to Storage techgroup and not 'Alliances 3rd-party products'.

Please change the techgroup to "storage" and re-submit the article to tech review.

Thanks and best regards,

Francois Lepretre
Knowledge Operations
Version: 0
Date: 2006-09-18
User Name: 95826
Action: Accept
Comment:
Version: 0
Date: 2006-08-02
User Name: 125432
Action: Approved
Comment: expert,

Please have a review for this Troubleshooting doc.

Thanks,
Willis Yao
Version: 0
Date: 2006-08-02
User Name: 125432
Action: Created
Comment:
Version: 0
Product_uuid
49311310-146b-11d7-870d-d17cb3a778c1|Brocade 12000 2 GB Switch
c5741319-fe44-11d9-a8ca-080020a9ed93|Brocade 48000 Director
24f55d62-6656-11d8-9e3a-080020a9ed93|Brocade SilkWorm 24000 Director

Attachments

This solution has no attachment