2同行回答
Cisco switch becomes unmanageable (VPD is corrupted or Invalid) - IBM BladeCenter
适用的国家和地区
Source
RETAIN tip: H204235
Symptom
A Cisco 3012 and 3110G switch Vital Product Data (VPD) becomes corrupt.
Case 1
The affected switch may become unmanageable, if an Advanced Management Module (AMM) is upgraded to AMM firmware version 3.54L - Build ID: BPET54L (or later) or experiences a failover while at that level (or later).
To determine if users have this issue, follow these steps:
Log in to AMM.
Navigate to I/O Module Tasks -> Configuration.
Select the slot for the Cisco switch.
The issue might exist if the panel reports the following message:
No management function is provided for this module type.
Follow these steps to confirm that users have this issue:
Collect AMM service data.
Open the service data .tgz file
Open the folder for primary First Failure Data Capture (FFDC).
Open the following file:
developer.txt
Search for the following string:
VPD Version:
The issue exists if there is service data such as either of the following two segments:
***** IOM SLOT [1]
(1) Present: Yes
(1) Type: Ethernet
(1) VPD Version: c304
***** IOM SLOT [7]
(7) Present: Yes
(7) Type: Ethernet
(7) VPD Version: d6a0
The VPD version should contain \'0104\' for the 3012 and 3110G switches. VPD Versions of \'c304\', \'d6a0\', or anything other than \'0104\' are incorrect and should be corrected using the steps outlined in the
Workaround section of this tip.
Case 2
There is an entry in the AMM event log such as the following:
I/O module X has incorrect VPD information.
And after the incorrect VPD message there is not a message such as the following:
Recovery I/O module X has incorrect VPD information.
Notes: In this case the VPD has become corrupted, but the AMM has cached the old VPD information. In this case, the AMM will report that the switch is manageable and service data will show the correct VPD.
Only after the AMM restarts will users see incorrect VPD in the service data and the AMM reporting the switch as unmanageable.
For either case, use the recovery steps outlined in the Workaround section of this tip.
In the event that it is not known which offset to use with the script, run the script first using offset \'0\'. If it fails :
rename the log file to ZZ-FIX_SWITCH_VPD_LOG-fail.txt
rerun the script using offset \'2\'
collect all logs and remove both script and logs from customer environment before departure
Affected Configurations
The system may be any of the following IBM servers:
BladeCenter E, type 8677, any model
BladeCenter H, type 7989, any model
BladeCenter H, type 8852, any model
BladeCenter HT (AC), type 8750, any model
BladeCenter HT (DC), type 8740, any model
BladeCenter S, type 1948, any model
BladeCenter S, type 7779, any model
BladeCenter S, type 8886, any model
The system is configured with one or more of the following IBM Options:
Cisco Catalyst Switch Module 3012 for IBM BladeCenter, Option part number 43W4395, any replacement part number (FRU)
Cisco Catalyst Switch Module 3110G for IBM BladeCenter, Option part number 41Y8523, any replacement part number (FRU)
Cisco Catalyst Switch Module 3110X for IBM BladeCenter, Option part number 41Y8522, any replacement part number (FRU)
This tip is not software specific.
The system has the symptom described above.
Workaround
The preferred recovery is to run a script that resolves the issue by correcting the VPD. If the user agrees, provide them with the script and run it on the affected system.
Where allowed, IBM service personnel may run the script locally for the user.
Note: Service Support Representatives (SSRs) in the United States (US) are not allowed to install non-IBM Standard Software Installer (ISSI) applications on their workstations, so installing Perl or the Perl script may not be possible.
If users allows remote access, then the case can be escalated to service personnel, such as Product Field Engineering (PFE) or Product Engineering (PE), who can run the script remotely. If no alternative is found, then a replacement switch will be sent as a Customer Replaceable Unit (CRU).
Any listed switch can be fixed by running the attached script.
Rename the script to \'fix-vpd-ssh.pl.\'
Note: The script uses Perl and Secure Shell (SSH) to perform actions, so the system that runs the script should have:
Perl installed
SSH client command line installed
SSH enabled on AMM
Using user supplied arguments, the script logs into the chassis, verifies that the switch has the corrupt VPD, and corrects the VPD.
Note: The script uses Perl and Secure Shell (SSH) to perform actions, so the system that runs the script must have:
Perl installed
SSH client command line installed
SSH enabled on AMM
To invoke the script, download it to machine with Perl/SSH client, rename the script to \'fix-VPD-SSH.pl\', then invoke the script as follows:
perl fix-vpd-ssh.pl AMM_IP_ADDRESS AMM_LOGIN_ID AFFECTED_SWITCH_BAY OFFSET CHASSIS_TYPE
where:
AMM_IP_ADDRESS - is the IP address of AMM
AMM_LOGIN_ID - is the login ID for AMM, for example, \'USERID\'
AFFECTED_SWITCH_BAY - is the bay with switch having incorrect VPD.
Valid values are 1, 2, 3, 4, 7, 8, 9, 10
OFFSET - is offset to area of corruption. Valid values are:
0 - if VPD version shows \'c304\'
2 - if VPD version shows \'d6a0\'
CHASSIS_TYPE - is type of chassis (if 8677).
Valid values are: \'BCE\' and \'BCER\' (default=BCE)
Note: 8677 has two (2) midplane types (BCE and BCER). If the script fails running on 8677 with default type (BCE), try running with \'type=BCER.\'
For example,
if AMM at \'9.42.161.57\', the AMM user ID is \'USERID\', and the switch in bay 1 has corruption with VPD version c304:
perl fix-VPD-SSH.pl 9.42.161.57 USERID 1 1 0
Users will be prompted up to four (4) times for the AMM password. After the script is run, the AMM will need to be restarted to see switch VPD properly.
Additional information :
The script now logs information to a file. This file will be created in the directory from which the script was run and will have the name of
ZZ-FIX_SWITCH_VPD_LOG.txt
If for any reason the script does not work with offset \'0\' or offset \'2\', include all log files generated by this script when escalating to PE.
Cisco has updated the switch firmware to block all ability for the management module to update VPD on the switch itself in firmware 15.0.2 SE5. Therefore, if corrupted VPD is found on switches that are now running at this level, the firmware will have to be downgraded (tested with 12.2.55 SE4), then the script can be run to fix the VPD, and then it can be reflashed to 15.0.2 SE5 or later after the VPD is corrected.
Additional Information
For the PE team:
The switch can be fixed by running the script provided in the Workaround section. The script is safe, but it is a
non-compiled clear-text perl script that provides access to powerful debug commands, which could cause failures when used incorrectly. The script is to be used as is and not modified. Any tampering with the script will void all support of the script and possibly the hardware it was used on.
Note: The script works in all IBM BladeCenter chassis. If the script fails for any reason, escalate to PE and be sure to include the log file from the failed run in the Problem Management Report (PMR). The log file should be located in the directory from which the script was run, and should have the name :
ZZ-FIX_SWITCH_VPD_LOG.txt.
Note: Download the script before each use to ensure the latest version of the script is utilized (download from the Attachments section of this Tip). Currently the most recent version of the script is 2014-05-12-fix-vpd-ssh.pl.
http://download.boulder.ibm.com/ibmdl/pub/systems/support/system_x_pdf/2014-05-12-fix-vpd-ssh.pl
收起