Sun System Handbook - ISO 3.4 June 2011 Internal/Partner Edition | |||
|
|
Solution Type Problem Resolution Sure Solution 1004618.1 : Sun Fire[TM] 12K/15K: frad daemon reports "Ignoring CRC error" or "TH Segment CRC is wrong on DIMM"
PreviouslyPublishedAs 206404
Applies to:Sun Fire 12K ServerSun Fire 15K Server All Platforms SymptomsThe frad daemon reports "Ignoring CRC error" or "TH Segment CRC is wrong on DIMM" errors into error log file in /var/opt/SUNWSMS/SMS1.2/adm/platform.Every four hours, a record of powered-on components is written to the SEEPROM. You may see hourly errors describing "Ignoring CRC error on DIMM with known bad CRC on DIMM" or "TH Segment CRC is wrong on DIMM". If you have one of the older DIMMS (for example, part number 501-5401-01 DIMM) that has the older CRCs in the SEEPROM, you may see platform error messages similar to those shown below (in file /var/opt/SUNWSMS/SMS1.2/adm/platform): Jul 25 10:14:29 2002 Q8SJ3 frad[442]: [9926 594352968222058 NOTICE SeepromInfoPro.cc 2715] Ignoring CRC error on DIMM with known bad CRC on DIMM at SB9/P1/B0/D3 Or you may see messages such as: Dec 17 08:18:14 2003 dwcoresc04-01 esmd[19577]: [1992 6553319427939517 ERR DynamicFru.cc 355] Failed to write the power event record, STILL_ON, of fru SB12/P2/B0/D0: rc=-3 CauseTechnically, there is a problem with the DIMM. The SEEPROM on SB9/P1/B0/D3 (first example) and SB12/P2/B0/D0 (second example) has a corrupt entry/segment. However, this is NOT an error in the memory area used by the domain. That's why you're not seeing CE ECCs or Rstops from the domain side.SolutionTo have the error messages cease, the DIMM must be replaced. However, because it has no impact on the health or operation of the domain, it can be done at the next scheduled maintenance window. You do not need to schedule a maintenance window for the sole purpose of replacing this DIMM.Product Sun Fire 15K Server Sun Fire 12K Server Internal Section Previously Published As 48156 This common issue can generate calls from customers who are concerned by the messages. As the doc shows, they should not be alarmed, and can wait to have service done until they can schedule service. Info suchas this should be available to Customers on MOS to prevent cases from being logged until they are ready to service it. Attachments This solution has no attachment |
||||||||||||
|