Paul, Thanks for the info. I believe it is OK to keep the memory module until we get a replacement.
Antonello Paul Dorian wrote: > > Hi Antonello, > > I have every reason to believe the pages are being retired. > > Look at fmdump -a it will show you pages retired. > > Sorry I don't know the mdb command but I'm sure you > could query and find out what memory has been retired > > HTH > > -Paul > > > Antonello Cruz wrote: >> One of our build servers (coupe.sfbay) reported a memory module " >> experiencing excessive correctable errors affecting large numbers of >> pages." >> >> The output of 'fmadm faulty' is: >> --------------- ------------------------------------ -------------- >> --------- >> TIME EVENT-ID MSG-ID SEVERITY >> --------------- ------------------------------------ -------------- >> --------- >> Nov 13 12:34:32 bd61220f-bd1c-45d0-9884-f4d1c8b94dc1 GMCA-8000-YN Major >> >> Fault class : fault.memory.generic-x86.dimm_ce >> Affects : >> hc://:product-id=Sun-Fire-X4240:chassis-id=0826QAS0AC:server-id=coupe/motherboard=0/chip=1/memory-controller=0/dram-channel=0/chip-select=0 >> >> >> faulted but still in service >> Problem in : >> hc://:product-id=Sun-Fire-X4240:chassis-id=0826QAS0AC:server-id=coupe/motherboard=0/chip=1/memory-controller=0/dram-channel=0/chip-select=0 >> >> >> faulty >> >> Description : A memory module is experiencing excessive correctable >> errors >> affecting large numbers of pages. Refer to >> http://sun.com/msg/GMCA-8000-YN for more information. >> >> Response : >> >> If Solaris is running on bare-metal (native) then >> affected memory >> pages associated with this memory module will be >> removed from >> service. >> >> Impact : >> >> Page retirement (where supported) is capped at a small >> fraction >> of memory to avoid performance impact. >> >> Action : Schedule a repair action to replace the memory module >> indicated >> by 'fmadm faulty'. While the errors are correctable in >> nature >> they may be a precursor to an uncorrectable error which >> will >> result in downtime. >> >> >> Lab support said it may take about a week to get the memory module to >> replace it. >> >> My question is if the faulty pages are not retired automatically by >> fma, how do I retire them while we wait for the memory module >> replacement? I would prefer not to remove the memory module, if >> possible. The machine is a x4240. >> >> Thanks, >> >> Antonello >> _______________________________________________ >> fm-discuss mailing list >> fm-discuss@opensolaris.org >> > _______________________________________________ fm-discuss mailing list fm-discuss@opensolaris.org
