Path: utzoo!attcan!uunet!lll-winken!lll-lcc!ames!xanth!nic.MR.NET!tank!mimsy!chris
From: chris@mimsy.UUCP (Chris Torek)
Newsgroups: comp.unix.wizards
Subject: Re: Help with uda errors
Message-ID: <15178@mimsy.UUCP>
Date: 29 Dec 88 05:54:50 GMT
References: <17846@adm.BRL.MIL> <15104@mimsy.UUCP> <26927@bu-cs.BU.EDU>
Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742
Lines: 26

In article <26927@bu-cs.BU.EDU> eap@bu-cs.BU.EDU (Eric Pearce) writes:
>Dec 20 17:14:32 bucsb vmunix: uda0: soft error, disk transfer error, 
>unit 2, grp 0x0, hdr 0xd4c27, event 0650

0650 is a 6-symbol ecc error (therefore correctable and corrected).
The 4.3BSD-tahoe driver decodes these things, printing something like

	uda0: soft error, disk transfer error: unit 2, lbn 871463:
	6 symbol ecc error (code A, subcode B)

The code and subcode are there in case DEC suddenly define new error codes.

>I did a level 0 dump of the entire disk.  DEC ran 'EVRLB' from the
>diag supervisor and then I restored the disk.  So far, no errors.

I cannot keep them straight, but presumably EVRLB forwards any marginal
sectors it finds.

>Was this the "correct" thing to do?

DEC have several `formatters' for forwarding bad sectors.  None of
them are capable of restoring an HDA to a virgin state, but at least
one of them (once called `rabads') can forward a sector by LBN.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163)
Domain:	chris@mimsy.umd.edu	Path:	uunet!mimsy!chris