chytil@vlsivie.tuwien.ac.at (Chytil Georg) (11/29/90)
I've just tried to invol our new 400t's SCSI-Disk w6:0 and after choosing option 1 (formating virgin physical volume) pressed <RETURN> twice, be it by accident or through a tossing (word?) keyboard. Be warned ! While this may happen quite easy it's a proper way to make Your disk unrecognisable to invol, salvol, mtvol and so on : 'invalid physical volume label','unable to read badspot list' ... Since I fear that our friendly HP-technician will take a time to find out about a remedy, I wonder if somebody else knows how to 'destuck'. I assume that rwvol may offer a chance, but I'm not that daring to invoke it without some hints. Now that I'm at it : Are You also experiencing lots of nasties in the MD of the 400t ? 'EX'ing a commandfile, e.g. salvol, leaving it and trying to 'EX' another just barks with strange error-messages in Hex. Is that why the reset-button is on the front ;-) ? Georg
krowitz@RICHTER.MIT.EDU (David Krowitz) (11/30/90)
Try using INVOL's option 7 to recreate a bad-block list. I don't have any HP 9000's, but if the FBS program is included (find-bad-spot) and if the version is similar to the latest versions on the SAU7/SAU8/SAU9 machines, then FBS can rebuild the bad-block list as it tests the disk drive. Earlier versions of FBS just typed out the numbers of the bad blocks on the console -- scrolling most of them off the top of the screen -:(, the newer version I've used is *much* nicer! Be forwarned, FBS takes *forever*! (after all, it has to write and re-read each block on the disk with several different data patterns) On my 155 MB DN4000 it runs for 2 to 3 hours per pass, and the time is dependent on the disk transfer rate and the size of the disk, not on the speed of the CPU. -- David Krowitz krowitz@richter.mit.edu (18.83.0.109) krowitz%richter.mit.edu@eddie.mit.edu krowitz%richter.mit.edu@mitvma.bitnet (in order of decreasing preference)
out117w@monu6.cc.monash.edu.au (y ng) (12/01/90)
In article <2185@tuvie.UUCP>, chytil@vlsivie.tuwien.ac.at (Chytil Georg) writes: > I've just tried to invol our new 400t's SCSI-Disk w6:0 and after choosing > option 1 (formating virgin physical volume) pressed <RETURN> twice, be it > by accident or through a tossing (word?) keyboard. > > Be warned ! While this may happen quite easy it's a proper way to make Your > disk unrecognisable to invol, salvol, mtvol and so on : 'invalid physical > volume label','unable to read badspot list' ... > We had similar experience when involing an external 660mb storage module connected to the SCSI port on the 400t running SR10.2. We contacted the local hp response centre and they could not figure out why and suggested that we need sr10.3 to support this new drive. Unfortunately 10.3 is not yet available in Australia so we decided to investigate further. After hours of experimenting with the invol we finally found the way to get out of this mess. If you initialise the bad spot list to null you can then proceed as usual. I presume SCSI drive manages its bad spot list internally? As far as the host computer is concerned the drive is bad spot free ??
chytil@vlsivie.tuwien.ac.at (Chytil Georg) (12/01/90)
As has been suggested on the net (Thanks !) reinitializing the badspot-list makes the drive involable ( sic ! ) again. Boy, am I relieved !!! A short hint in some invol-description could have saved me a lot of time and troubles, but it says "a string of Your choice", and according to Hopcroft/Ullman epsilon is a proper string, isn't it ? ;-) Georg
krowitz@RICHTER.MIT.EDU (David Krowitz) (12/03/90)
I believe you are correct. All the SCSI drives I have encountered so far manage their own bad spot lists internally. As far as the CPU was concerned (this is on own DN2500's), the drives are error-free. If you wish to check this, you can use the FBS (find-bad-spots) program located in your appropiate sau (stand-alone-utilities) directory (ie. /sau9/fbs for a DN2500, /sau8/fbs for a DN3000, /sau7/fbs for a DN 3500/4000/4500). This program will read and write all blocks on the disk with several data patterns looking for errors and will add any bad-spots it finds to the bad block list. Be forwarned: FBS takes several hours to run, and since it runs from the MD your machine will be down for the entire period. There is also a version of FBS in the /systest/ssr_util directory which will execute with your machine up and running, checking a disk that has not been mounted. I haven't used this on-line version, so I can't say whether or not it's any faster than the off-line version. -- David Krowitz krowitz@richter.mit.edu (18.83.0.109) krowitz%richter.mit.edu@eddie.mit.edu krowitz%richter.mit.edu@mitvma.bitnet (in order of decreasing preference)
sboyle@mentor.com (Sean Boyle x1542) (12/29/90)
In article <2185@tuvie.UUCP> chytil@vlsivie.tuwien.ac.at (Chytil Georg) writes: >I've just tried to invol our new 400t's SCSI-Disk w6:0 and after choosing >option 1 (formating virgin physical volume) pressed <RETURN> twice, be it >by accident or through a tossing (word?) keyboard. > >Be warned ! While this may happen quite easy it's a proper way to make Your >disk unrecognisable to invol, salvol, mtvol and so on : 'invalid physical >volume label','unable to read badspot list' ... > >Since I fear that our friendly HP-technician will take a time to find out >about a remedy, I wonder if somebody else knows how to 'destuck'. I >assume that rwvol may offer a chance, but I'm not that daring to invoke it >without some hints. .PP I suppose you could use rwvol in combination with the debugger, but that's rather involved. A better choice would be to use /systest/ssr_util/fixvol. .PP Once you are in, you can type in 'h' for help, but here is a basic runthrough. .nf (fv [p])> r pv ...blah blah blah... (fv [p])> edit pv pv> h -->Legal choices are: q up vers apollo[] name[] uid[] dtype b_per_vol s_per_trk t_per_cyl lv_list[] alv_list[] bs_daddr diag_daddr sect_st sect_size pre_comp b_per_pvol boot_daddr grp_size grp_my_n grp_nums[] grp_alg s_per_blk soft_sect ver_len version[] cookie pdaddr pv> name[] ( 8) name[0] == ( 0) : `I ( 9) name[1] == ( 0) : `' ( A) name[2] == ( 0) : `m ( B) name[3] == ( 0) : `_ ( C) name[4] == ( 0) : `s ( D) name[5] == ( 0) : `o ( E) name[6] == ( 0) : `_ ( F) name[7] == ( 0) : `f ( 10) name[8] == ( 0) : `o ( 11) name[9] == ( 0) : `r ( 12) name[A] == ( 0) : `g ( 13) name[B] == ( 0) : `e ( 14) name[C] == ( 0) : `t ( 15) name[D] == ( 0) : `f ( 16) name[E] == ( 0) : `u ( 17) name[F] == ( 0) : `l ( 18) name[10] == ( 0) : ( 19) name[11] == ( 0) : ( 1A) name[12] == ( 0) : ( 1B) name[13] == ( 0) : ( 1C) name[14] == ( 0) : ( 1D) name[15] == ( 0) : ( 1E) name[16] == ( 0) : ( 1F) name[17] == ( 0) : ( 20) name[18] == ( 0) : ( 21) name[19] == ( 0) : ( 22) name[1A] == ( 0) : ( 23) name[1B] == ( 0) : ( 24) name[1C] == ( 0) : ( 25) name[1D] == ( 0) : ( 26) name[1E] == ( 0) : ( 27) name[1F] == ( 0) : pv> q (fv [p])> show pv version: 0, apollo: , name: "I'm_so_forgetful" uid: 0.00000000 uid_$nil, built on 0 dtype: 0, blocks_per_pvol: 0 (0) sects_per_track: 0, tracks_per_cyl: 0 phys_badspot_daddr: 0, phys_diag_daddr: 0 phys_sector_start: 0, phys_sector_size: 0 (0), pre_comp: 0 (fv [p])> w . . . -- "There is a time to laugh and a time not to laugh and this is not one of them." Inspector Jacques Clouseau sboyle@mentor.com Mentor Graphics Corporation