myamin@cbnewsm.ATT.COM (m.yamin) (09/21/89)
Does anyone know what HDERR error flag 1 means? Here is a specimen line from unix.log: HDERR ST:51 EF:1 CL:4282 CH:4200 SN:4208 SC:4202 SDH:4224 DMACNT:FFFF DCRREG:94 MCRREG:8900 Tue Sep 19 20:58:09 1989 The disk is a new Seagate 251-1 controlled by the original WD1010. gdisk.h offers some interpretation of the status and error codes thus: /* * Status information for WD1010. */ #define W_ERR 0x0001 /* logical or of bits in the error register */ #define W_CIP 0x0002 /* 1= command in progress */ #define W_DRQ 0x0008 /* 1= data request outstanding */ #define W_SEKCMP 0x0010 /* 1= seek complete */ #define W_WRFLT 0x0020 /* 1= write fault */ #define W_RDY 0x0040 /* 1= ready,frozen on error until status read */ #define W_BUSY 0x0080 /* 1= WD1010 is accessing the disk */ /* * Error status information for the WD1010. */ #define W_BBD 0x0080 /* bad block detected (not used) */ #define W_CRC 0x0040 /* 1= CRC error */ #define W_ID 0x0010 /* 1= id not found */ #define W_ABC 0x0004 /* 1= aborted command */ #define W_TK 0x0002 /* 1= restore command did not find track zero */ So ST:51 means seek is complete, the register is frozen until the status is collected, and there is a bit set in the error register. EF:1, I suppose, means that's what's in the error register, but no interpretation is given for this bit (0x0001). Some days I get hosts of these, other days nary a one; it seems to depend on what mood the disk woke up in. (For example, yesterday I got enough so that unix.log cleared itself several times, and since booting today I haven't seen a single error, and I haven't touched the machine). The "errors" do not affect the operation or performance of the system in any way, do not result in wrong data, and do not result in audible re-seeks or recalibrations as EF:40's and EF:10's do; I know what those sound like, because I heard plenty of them on my old Miniscribe. EF:1 is the only error flag I've seen on this disk. M. Yamin myamin@cbnewsm.ATT.COM
jcm@mtunb.ATT.COM (was-John McMillan) (09/22/89)
In article <4387@cbnewsm.ATT.COM> myamin@cbnewsm.ATT.COM (m.yamin) writes: >Does anyone know what HDERR error flag 1 means? Here is a specimen >line from unix.log: > >HDERR ST:51 EF:1 CL:4282 CH:4200 SN:4208 SC:4202 SDH:4224 DMACNT:FFFF >DCRREG:94 MCRREG:8900 Tue Sep 19 20:58:09 1989 The error means the Data Address Mark was not found. The DAM is part of the meta-information written during formatting. It should never be re-written until the next formatting. A marginal recording of this mark, or a marginal bit of surface damage might make this an intermittent error. You HAVE NOT mentioned if these errors are logged as being on the same cylinder/sector, however, and if there is only one damaged sector it's quite possible you could on access it intermittently. Errors are logged BEFORE they are total. Preferably there'd be a retry-count message alongside, but that code was misplaced slightly -- I believe -- causing it to not be logged under certain circumstances. It is quite possible you are just having to perform a re-read and are therefore NOT losing anything but the spin time. The message would disappear if the damaged sector were added to your bad block table -- or, probably, if you re-format your disk. john mcmillan -- att!mtunb!jcm -- speaking for self, only...