Red Hat Bugzilla – Bug 1048207
dmidecode in rhel6 does not fully support SMBIOS 2.8 on HP gen8
Last modified: 2017-10-19 04:10:26 EDT
Description of problem: With the latest bios version (2013.11.14) on HP BL460c Gen8 blade servers dmidecode reports: # SMBIOS implementations newer than version 2.7 are not # fully supported by this version of dmidecode at the start of all output. Otherwise the output looks to be correct. This additional warning message is enough to break scripts expecting a single line of output. Version-Release number of selected component (if applicable): dmidecode-2.11-2.el6_1.x86_64 How reproducible: Every time Steps to Reproduce: 1. run "dmidecode -s system-product-name" or similar 2. 3. Actual results: # SMBIOS implementations newer than version 2.7 are not # fully supported by this version of dmidecode ProLiant BL460c Gen8 Expected results: ProLiant BL460c Gen8 Additional info: The full dmidecode ouput includes the line "SMBIOS 2.8 present". The 2013.09.18 version of the bios does not produce this message.
This has also been confirmed on DL360p Gen8 and DL380p Gen8 servers with the 2013.11.14 bios
This has also been reported to me on RHEL5 with dmidecode-2.11-1.el5.x86_64
Chris, could you attach the binary dump of this system to bugzilla? see dmidecode's --dump-bin option. thanks!
We have tried compiling the source available from http://download.savannah.gnu.org/releases/dmidecode/dmidecode-2.12.tar.gz Which gives us the same message (but reporting a different dmidecode version): # dmidecode 2.12 SMBIOS 2.8 present. # SMBIOS implementations newer than version 2.7 are not # fully supported by this version of dmidecode. We have also compiled the cvs version of the source and that works correctly: # dmidecode 2.12 SMBIOS 2.8 present. 192 structures occupying 6294 bytes [etc etc] (both compilations on the same RHEL6.4 DL380p) Is it posible to get the CVS version into RHEL 5&6 as an errata?
If there is a demand. Let's try, will see what PMs say. :) You may want to clone this bug for rhel5 too.
This is impacting us as well, dmidecode scripts are commonly used to read the vendor serial number of hardware for automated support tools, or in our cases as a fingerprint/primary key for the server
FYI, this change went into RHEL5 (bug 1048920).