acphssrw@csuna.UUCP (Stephen R. Walton) (06/17/88)
In article <2425@amiga.UUCP> jimm@cloyd.UUCP (Jim Mackraz) writes: > [other ideas about speeding up booting deleted] > [under V1.2]: > > Since Execute doesn't use T: for it's temp file, it might > use SYS:T. In that case, assign sys: before you run your > hard disk script. > > If I'm wrong on that, it uses :T, so cd to your fast disk > before executing the hard disk script. Yes, 1.2 Execute always uses :T for its temporary. However, there is a real problem with doing this, as I found out: it results in a very slow revalidation of your entire hard disk if your Startup-Sequence does a LoadWB. I can email the details (or you can read 'em on B$X), but briefly, if you do something like: assign (everything to dh0:) cd dh0: execute s:Startup-Sequence on your boot floppy, Don't Put LoadWB In The DH0: Startup-Sequence. Put it in the one on the floppy or do it by hand after bootup. Stephen Walton, representing myself swalton@solar.stanford.edu Cal State, Northridge rckg01m@calstate.BITNET
riley@batcomputer.tn.cornell.edu (Daniel S. Riley) (06/17/88)
In article <1249@csuna.UUCP> swalton@solar.stanford.edu (Stephen R. Walton) writes: >Yes, 1.2 Execute always uses :T for its temporary. However, there is >a real problem with doing this, as I found out: it results in a very >slow revalidation of your entire hard disk if your Startup-Sequence >does a LoadWB. I can email the details (or you can read 'em on B$X), I haven't read the thread on BIX, but this seems a little extreme. As I understand the situation, LoadWB does something equivalent to pulling a disk out and sticking it back in (ACTION_something, diskchange or inhibit, I guess), so you just need to avoid anything that you wouldn't want to be happening when you pull out a disk. Most people get burnt by doing something like "date >now" followed by loadwb, which gets them because of the 3 second DOS delay in updating the disk. I would think that as long as you wait longer than 3 seconds after the last write, and don't have any files open for write (like .key temporary files from execute), you should have no problems. This certainly seems to be true on floppies, and I would expect it to be the same with a hard disk. Two questions: 1. Am I missing something here? (I'm about to get a hard disk, so I'm interested. 2. Why does LoadWB do whatever it does? -Dan Riley (dsr@lns61.tn.cornell.edu, dsr@crnlns.bitnet) -Wilson Lab, Cornell U.
lphillips@lpami.van-bc.UUCP (Larry Phillips) (06/19/88)
In <1249@csuna.UUCP>, acphssrw@csuna.UUCP (Stephen R. Walton) writes: > [some stuff about speeding up boots and the ( T: | :T ) directory] Better still, don't use EXECUTE at all (or filezap the :T to T: if you must), but instead use a technique that doesn't require a temporary file. My startup-sequence looks something like this: binddrivers ; use mount if applicable DH0:NewCLI from DH0:HDStart DH0:EndCLI >NIL: The rest of the startup sequence is, of course contained in the file DH0:HDStart, which is on a small, Old File System partition. By small, I mean it's 1 cylinder long on my Seagate ST225, and contains just what's needed to get the FFS up and running on the other partition. If you're using the Old File System or FFS with a drive that is MOUNTed rather than BINDDRIVERed, you can go directly for the commands in the DHn:C and for the script file in DHn:S -larry -- If all the MSDos machines were laid end to end, they still wouldn't be as fun as a single Amiga. +----------------------------------------------------------------+ | // Larry Phillips | | \X/ {ihnp4!alberta!ubc-vision,uunet}!van-bc!lpami!lphillips | | COMPUSERVE: 76703,4322 | +----------------------------------------------------------------+
andy@cbmvax.UUCP (Andy Finkel) (06/21/88)
In article <5200@batcomputer.tn.cornell.edu> riley@tcgould.tn.cornell.edu (Daniel S. Riley) writes: >In article <1249@csuna.UUCP> swalton@solar.stanford.edu (Stephen R. Walton) writes: >>Yes, 1.2 Execute always uses :T for its temporary. However, there is >>a real problem with doing this, as I found out: it results in a very >>slow revalidation of your entire hard disk if your Startup-Sequence Actually, its the normal speed revalidation :-) >>does a LoadWB. I can email the details (or you can read 'em on B$X), > >I haven't read the thread on BIX, but this seems a little extreme. As >I understand the situation, LoadWB does something equivalent to pulling >a disk out and sticking it back in (ACTION_something, diskchange or >inhibit, I guess), so you just need to avoid anything that you wouldn't The reason Workbench does an ACTION_INHIBIT on all the drives is to identify the disk devices on the DOS device list. Only disk devices are expected to respond to the ACTION_INHIBIT packet. Under the old file system, this created a window of vunerability. Inhibit always succeeded...even if there were outstanding locks. (Fast File System can fail the ACTION_INHIBIT packet, btw). So, a WAIT after the LoadWb is probably a pretty good thing to do, if you plan on creating files in your startup-sequence. -- andy finkel {ihnp4|seismo|allegra}!cbmvax!andy Commodore-Amiga, Inc. "Un*x is the answer, but only if you phrase the question very carefully. Any expressed opinions are mine; but feel free to share. I disclaim all responsibilities, all shapes, all sizes, all colors.