larserio@IFI.UIO.NO (LarsErikOsterud) (12/18/90)
I have a BIG problem with THE VAULT hard disk backup !!! It won't work at all together with the DCACHE hard disk cache program. When DCACHE is installed and active VAULT works VEEERYYYY SLOOWLYYY and when it finally is finished ALL THE BACKUP DISKS ARE BLANKS !!! NO other programs has problems with DCACHE and if I am going to support THE VAULT it MUST work 100% together with DCACHE first... Does anyone know what's the problem and how to get passed it ??? Lars-Erik / ABK-BBS +47 2132659 / ____ ______ ________________________ Osterud / larserio@ifi.uio.no / /___ / The norwegian ST __________/ ______________________/ ____/ / Klubben, user association
ljdickey@watmath.waterloo.edu (L.J.Dickey) (12/19/90)
In article <CMM.0.88.661488577.larserio@kvart.ifi.uio.no> larserio@ifi.uio.no writes: >I have a BIG problem with THE VAULT hard disk backup !!! >It won't work at all together with the DCACHE hard disk cache program. Remember Turtle? Turtle insists that it be in control of everything. Its own ram disk, everything. I wonder of the writer of The Vault made the same assumptions, but did not build in the checks to make sure.
fischer-michael@cs.yale.edu (Michael Fischer) (12/19/90)
In article <1990Dec19.132054.28228@watmath.waterloo.edu> ljdickey@watmath.waterloo.edu (L.J.Dickey) writes: >In article <CMM.0.88.661488577.larserio@kvart.ifi.uio.no> > larserio@ifi.uio.no writes: >>I have a BIG problem with THE VAULT hard disk backup !!! >>It won't work at all together with the DCACHE hard disk cache program. > >Remember Turtle? Turtle insists that it be in control of everything. >Its own ram disk, everything. I wonder of the writer of The Vault >made the same assumptions, but did not build in the checks to make sure. Please, keep your cool and don't start speculations flying on the net. All that Lars reported is that he has a problem running THE VAULT and DCACHE together. We don't know what the problem is, but we are trying to find it. Meanwhile, don't run them together. They both apparently work fine for most people most of the time when used separately. TURTLE, too, seems to work well for many people, and I don't understand your implicit criticism of it. The problem with the VAULT/DCACHE combination could lie with either program, and it is not necessarily due to one of the writers making "assumptions"---it could be due to a plain old ordinary unintentional bug: accidently overwriting memory not belonging to the program, using a signed int someplace instead of an unsigned, a problem with the compiler or a library routine, etc. -- ================================================== | Michael Fischer <fischer-michael@cs.yale.edu> | ==================================================
fischer-michael@cs.yale.edu (Michael Fischer) (12/21/90)
In article <CMM.0.88.661488577.larserio@kvart.ifi.uio.no> larserio@ifi.uio.no writes: >I have a BIG problem with THE VAULT hard disk backup !!! >It won't work at all together with the DCACHE hard disk cache program. > >When DCACHE is installed and active VAULT works VEEERYYYY SLOOWLYYY >and when it finally is finished ALL THE BACKUP DISKS ARE BLANKS !!! I found the problem! It is with DCACHE, not The Vault. The problem has to do with the way DCACHE installs itself in the chain of rwabs handlers. Whenever DCACHE wants to call the old handler (for example, to pass through an rwabs call for a drive that it is not caching), instead of just passing control to the old handler, it temporarily reinstalls the old handler at the head of the rwabs chain and then traps (recursively) to the bios. The effect of this is to wipe out any rwabs handler that is installed after DCACHE. The Vault has its own caching scheme and installs its own rwabs handler. DCACHE immediately wipes it out, causing The Vault loses control of its cache. Files then go directly to floppy instead of to The Vault's cache. When the cache is finally flushed to disk, there is nothing there and the disk ends up empty. >NO other programs has problems with DCACHE This is not true. No program that installs an rwabs handler will work if run after DCACHE is installed. In particular, this applies to ramdisks. If the ramdisk is first in the \AUTO folder, then everything is fine, but if DCACHE is first, you'll find that the ramdisk doesn't work. I have attempted to contact the author of DCACHE to see if he can fix this problem with DCACHE. Until then, users of DCACHE should aware of the problems it can cause with other programs. -- ================================================== | Michael Fischer <fischer-michael@cs.yale.edu> | ==================================================
ljdickey@watmath.waterloo.edu (L.J.Dickey) (12/24/90)
In article <27875@cs.yale.edu> fischer-michael@cs.yale.edu (Michael Fischer) writes: >In article <1990Dec19.132054.28228@watmath.waterloo.edu> ljdickey@watmath.waterloo.edu (L.J.Dickey) writes: ... >>Remember Turtle? Turtle insists that it be in control of everything. ... >Please, keep your cool and don't start speculations flying on the net. >All that Lars reported is that he has a problem running THE VAULT and >DCACHE together. We don't know what the problem is, but we are trying >to find it. Meanwhile, don't run them together. They both apparently >work fine for most people most of the time when used separately. >TURTLE, too, seems to work well for many people, and I don't >understand your implicit criticism of it. Sorry. This was not intended to be a criticism of either VAULT or DCACHE. I am sure both are fine programs, but I have used only one, with no complaints. My criticism, if there was one, was implicit indeed, and would be more precisely aimed at the absence of a central direction or control in the development of software. I suspect that many users have found instances, as I have, of software, each one fine by itself, which clash with each other, because of assumptions made by the author. In some case *I assume* that these programmer assumptions include making use of so called "illegal" locations in the system. I do not wish to impune the integrity of any author here. I am sure that each does the best that she/he can with what is at hand, and I am grateful to a number of authors of fine software for the work that they have done. Again, sorry. -- Leroy J. Dickey, Faculty of Mathematics, U of Waterloo, Canada N2L 3G1 Internet: ljdickey@watmath.waterloo.edu UUCP: ljdickey@watmath.UUCP ..!uunet!watmath!ljdickey X.400: ljdickey@watmath.UWaterloo.ca BITNET: ljdickey@watdcs.bitnet
pegram@kira.UUCP (Robert B. Pegram) (12/28/90)
(L.J.Dickey) writes (eariler): > .... >>>Remember Turtle? Turtle insists that it be in control of everything. > .... > Then, in article <27875@cs.yale.edu> fischer-michael@cs.yale.edu (Michael Fischer) writes: >>Please, keep your cool and don't start speculations flying on the net. >>All that Lars reported is that he has a problem running THE VAULT and >>DCACHE together. We don't know what the problem is, but we are trying >>to find it. Meanwhile, don't run them together. They both apparently >>work fine for most people most of the time when used separately. >>TURTLE, too, seems to work well for many people, and I don't >>understand your implicit criticism of it. From article <1990Dec24.034049.21799@watmath.waterloo.edu>, by ljdickey@watmath.waterloo.edu (L.J.Dickey) replies: > Sorry. This was not intended to be a criticism of either VAULT or DCACHE. > I am sure both are fine programs, but I have used only one, with no > complaints. My criticism, if there was one, was implicit indeed, and > would be more precisely aimed at the absence of a central direction or > control in the development of software. I suspect that many users have > found instances, as I have, of software, each one fine by itself, which > clash with each other, because of assumptions made by the author. ....(small omission) > I am sure that each > does the best that she/he can with what is at hand, and I am grateful > to a number of authors of fine software for the work that they have > done. > Again, sorry. > -- > Leroy J. Dickey, Faculty of Mathematics, U of Waterloo, Canada N2L 3G1 > Internet: ljdickey@watmath.waterloo.edu > UUCP: ljdickey@watmath.UUCP ..!uunet!watmath!ljdickey > X.400: ljdickey@watmath.UWaterloo.ca > BITNET: ljdickey@watdcs.bitnet I am tired of conflicting programs too. Those reviews that people (myself included) should be writing for c.s.a.st and atari.archive should include compatability testing. I would also like to read reviews of peoples' working systems that *do*not*conflict! On a related topic, is it too much to ask for people to write single purpose programs? The glut of "everything but the kitchen sink" stuff means I can't pick the best, or I have to waste precious memory with DAs that duplicate functions. just my expensive $0.02, 8-) Bob Pegram pegram@griffin.uvm.edu or ...!uvm-gen!pegram