home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!xenitec!news
- From: belal@sco.COM (Bela Lubkin)
- Newsgroups: biz.sco.general
- Subject: boot-time delay, Re: Two things....
- Message-ID: <9212292327.aa06350@scoke.sco.COM>
- Date: 30 Dec 92 07:27:13 GMT
- Sender: news@xenitec.on.ca (xenitec.on.ca News Administrator)
- Organization: [resent by] The SCOGEN gateway and Propagation Society
- Lines: 38
- Resent-From: mmdf@xenitec.on.ca
- Submit-To: scogen@xenitec.on.ca
- Precedence: bulk
-
- Michael B Garrett wrote:
-
- > Also: Has anyone ever seen an ODT system that took about 1 minute to start
- > reading the .text and .data files on bootup? Mine waits at least that long
- > after you hit "enter" at the "Boot:" prompt. Since it has done this since
- > ODT 1.0, I assumed it was just a SCSI kinda thing, but my last client
- > got SCSI also, and it doesn't do it there. Both machines are EISA 486/33,
- > mine is a Bustek 742A and the client's is a 1740A.
- >
- > Is this slowness indicative of some problem I'm unaware of, or is it
- > just because the disk is slower (Maxtor 660MB) ?
-
- Are you sure this behavior dates back so far? (I sort of doubt it since
- at the time that ODT 1.0 was current, 486/33 EISA machines did not
- exist). I wonder because a delay at exactly that point in the boot
- process was introduced with SCO UNIX 3.2v4.0 and ODT 2.0. The delay is
- indicative of the boot program being extremely careful about caches.
- ODT 1.1 and UNIX 3.2v2 had problems on 486 machines because /boot could
- get confused by the 486 internal cache. The new code is "paranoid" and
- does (slow) cache-flush instructions while sizing memory.
-
- In a later release (I'm not sure which release this will appear in), the
- delay will be user-selectable (you can choose to have it skip the cache
- flushes, and *hope* it's safe); and, instead of sitting there looking
- like it hung, it will print dots:
-
- Boot
- :
-
- Sizing memory
- .................
- Loading hd(40)unix .text
- ................................
-
- Anyway, under 3.2v4.x and ODT 2.0, the delay is normal (and varies
- depending on CPU speed, cache design etc.)
-
- >Bela<
-