mph@rover.UUCP (Mark Huth) (06/25/88)
Okay, I know that this has been answered before, but what is the magic sequence required to boot a hard drive system and load Workbench from the hard drive? What I want to do is have the floppy sequence be as short as possible. Currently, the startup-sequence on df0: is: system/FastMemFirst BindDrivers execute dh2:s/startup-sequence dh2:s/startup-sequence: cd dh2: execute dh2:s/startup-assigns setmap usa1 dmouse rslclock Loadwb shell from s:shell-startup The startup-assigns assigns sys: to dh2: and all the rest of the logical devices found when a floppy Workbench boot has the assign command execute (ie, c: s: l: devs: fonts:, etc.). Also, the startup-assigns file sets the path to include current directory, sys:system, sys:utilities, ram:, and sys:. When the script gets to the loadwb command I get a requestor telling me that the disk isn't validated yet, and of course, the disk validator is running (for about 10 -15 minutes - 40 meg disk). After the validator finishes, workbench loads and the script completes. How do I get this script to work without having the validator hold things up forever. It doesn't happen if I just boot the cli and shell, or if the loadwb command is located before all the assigns to the hard disk (ie it is loaded from the boot floppy.) Thanks, Mark Huth
rss@ece-csc.UUCP (ML) (06/25/88)
In a previous article mph@rover.UUCP (Mark Huth) wrote: >Okay, I know that this has been answered before, but what is the magic >sequence required to boot a hard drive system and load Workbench from >the hard drive? What I want to do is have the floppy sequence be as >short as possible.... ... >When the script gets to the loadwb command I get a requestor telling >me that the disk isn't validated yet, and of course, the disk >validator is running (for about 10 -15 minutes - 40 meg disk). After >the validator finishes, workbench loads and the script completes. > >How do I get this script to work without having the validator hold >things up forever. It doesn't happen if I just boot the cli and >shell, or if the loadwb command is located before all the assigns to >the hard disk (ie it is loaded from the boot floppy.) As I understand it, the guilty culprit here is the "execute" command which is writing a temporary file on your hard drive which is eventually the cause of the tremendous delays incurred by the disk validator when you issue the LoadWB command. Rather than using "execute"; do a "NewCli FROM file". Here's my startup sequence from df0:s ************************* cut here **************************************** Wait 5 ; Allow hard drive to come up to speed Mount DH0: ; Bring it online Assign LIBS: DH0:Libs ; Assignments to LIBS and L are so that Assign L: DH0:L ; I can run the ARP commands on DH0. Assign DEVS: DH0:Devs ; To get at MountList before using ConMan Assign C: DH0:C ConMan > NIL: -w NewWsh "CON:0/10/640/190/Master Wshell/c" From DH0:S/BootSequence EndCli > NIL: ************************* snip snip *************************************** The initial "Wait 5" I put there because sometimes I found the "Mount" command would fail when I first turned on the machine, apparently because the drive wasn't ready yet. The initial ASSIGN statements (except perhaps C: to DH0:C) aren't necessary if you are running your machine with standard Amiga commands; they're only there in my sequence because (a) I have the ARP commands installed on my hard drive, and they need a run-time library in LIBS:, and (b) I use ConMan, which has a handler in L:, plus an entry in DEVS:MountList. Of course, like you, the bulk of my startup stuff is on the hard drive (in DH0:S/Bootsequence), where I do all the usual sorts of things (load AREXX, AmiCron, DMouse [love that Dmouse!], CD DH0:, LoadWB, etc) and all works quickly. Hope this helps, Mark Lanzo thru ..mcnc!ece-csc!rss
rodger@hpdml93.HP.COM (rodger anderson) (06/29/88)
>/ hpdml93:comp.sys.amiga / mph@rover.UUCP (Mark Huth) / 11:04 am Jun 24, 1988 / >Okay, I know that this has been answered before, but what is the magic >sequence required to boot a hard drive system and load Workbench from >the hard drive? What I want to do is have the floppy sequence be as >short as possible. Currently, the startup-sequence on df0: is: >system/FastMemFirst >BindDrivers cd ram: ; You need this here for the execute command. >execute dh2:s/startup-sequence >dh2:s/startup-sequence: >cd dh2: ^^^^^^^ take this out, stay in ram. >execute dh2:s/startup-assigns >setmap usa1 >dmouse >rslclock >Loadwb >shell from s:shell-startup When you use the dos execute command, it creates a file in :T called command-<something>. If this is is on the floppy, it slows things down as the script is read from the floppy. If this is from the hard disk, you may have the problems you describe. >When the script gets to the loadwb command I get a requestor telling >me that the disk isn't validated yet, and of course, the disk >validator is running (for about 10 -15 minutes - 40 meg disk). After >the validator finishes, workbench loads and the script completes. >How do I get this script to work without having the validator hold >things up forever. It doesn't happen if I just boot the cli and >shell, or if the loadwb command is located before all the assigns to >the hard disk (ie it is loaded from the boot floppy.) I have heard you can also use newcli instead of execute, and eliminate the creation of a temporary file, but I haven't tried this yet. >Thanks, Your welcome. >Mark Huth >---------- -- Rodger Anderson (rodger@hpdml93)
mlelstv@faui44.informatik.uni-erlangen.de (Michael van Elst ) (06/29/88)
In article <782@rover.UUCP> mph@rover.UUCP (Mark Huth) writes: >. >. >When the script gets to the loadwb command I get a requestor telling >me that the disk isn't validated yet, and of course, the disk >validator is running (for about 10 -15 minutes - 40 meg disk). After >the validator finishes, workbench loads and the script completes. > >How do I get this script to work without having the validator hold >things up forever. It doesn't happen if I just boot the cli and >shell, or if the loadwb command is located before all the assigns to >the hard disk (ie it is loaded from the boot floppy.) > >Thanks, > >Mark Huth I got the same problems when I tried to use a startup-sequence similar to this: >> Mount DH0: >> DH0:c/Execute DH0:s/startup-sequence The problem is that in order to run a nested script an intermediate file 'Command-T-n' is generated in the :T directory. When the script execution reaches a LoadWB the workbench program sends (ACTION_INHIBIT) to all filesystem tasks. Thus, a filesystem that has been written to cannot update its rootblock, i.e. it is not validated. The first solution is to use RAM: as the directory for the intermediate file. Disadvatage: the Ram-Handler has to be loaded from disk. The second solution (I have used for 3 months) is to >> DH0:c/RUN DH0:c/Execute DH0:s/startup-sequence Disadvantage: the second CLI Task cannot accept any input. The third solution (and the best one): Wait for the rootblock to get updated before you start Workbench. This requires a 'WAIT 3' in your second startup-sequence (that one on your harddisk). If you have more drives you may have to increase the time you wait. -- Michael van Elst E-mail: UUCP: ...seismo!unido!fauern!faui10!mlelstv E-mail: UUCP: ...uunet!unido!fauern!faui10!mlelstv <- when seismo ceases operation
mph@rover.UUCP (Mark Huth) (07/06/88)
In article <3673@ece-csc.UUCP> rss@ece-csc.UUCP (ML) writes: >In a previous article mph@rover.UUCP (Mark Huth) wrote: >... >>When the script gets to the loadwb command I get a requestor telling >>me that the disk isn't validated yet, and of course, the disk > >As I understand it, the guilty culprit here is the "execute" command >which is writing a temporary file on your hard drive which is eventually Thanks, Mark Lanzo I have actually been told that a WAIT 5 after the execute of the hd startup sequence would fix the problem. It does. I've simply added WAIT 5 between Execute and LoadWb. This cures the validator problem. Mark Huth