fx/disks/3.2 upgrade

Mark Bradley markb at denali.sgi.com
Wed Dec 20 11:30:20 AEST 1989


In article <17280030 at acf4.NYU.EDU>, wood at acf4.NYU.EDU (David Wood) writes:
> 
> 		We recently upgraded to 3.2 on our 4D/80GT.  Last
> 	week we began noticing unrecovered disk errors on the
> 	disk with / and /usr.  So I read the man page for fx
> 	and used fx to find the bad blocks.  My mistake was using
> 	fx to forward the bad blocks while the system was up
> 	and running.  The resulting sympton was that we got disk 
> 	I/O errors during 'find ...' .  I double checked the man page 
> 	and nowhere does it say you can't do this.  BE WARNED that you 
> 	are supposed to use the stand alone version of fx when the
> 	system is down of course.  Incidentally, we have an ESDI drive.
> 		While discussing this with SGI, one of there field
> 	service guys told me that when upgrading to 3.2, you are
> 	supposed to re-format your disks.  "Really, it didn't say that
> 	anywhere in the installation guide", says I.   He said
> 	it was only recently discovered that this was necessary.
> 	Has anyone else heard anything like this?
> 
No, no, no.  One does not need to reformat when upgrading s/w.  An mkfs
is nice once in a great while to de-frag the filesystem in a reload from
tape, but you do NOT need to reformat your drive when upgrading s/w.  Key
word is tape--make sure you are backed up on tape first.

You CAN forward bad blocks using the run-time version of fx, but of course
in trying to read data from a bad track (you know it's become bad because 
you get errors trying to read it, right?) , you will get errors.  You have
to read the data to forward it, naturally.

						markb
--
Mark Bradley				"Faster, faster, until the thrill of
I/O Subsystems				 speed overcomes the fear of death."
Silicon Graphics Computer Systems
Mountain View, CA 94039-7311		     ---Hunter S. Thompson

********************************************************************************
* Disclaimer:  Anything I say is my opinion.  If someone else wants to use it, *
*             it will cost...						       *
********************************************************************************



More information about the Comp.sys.sgi mailing list