sahayman@iuvax.cs.indiana.edu (Steve Hayman) (12/15/88)
Timothy R. Giebelhaus of Apollo writes: >You can get much more sophisticated by manipulating the backup_history files. >For example, you can get muliple levels of dumps by saving multiple >backup_history files. Though this is more complicated, it can be handled >through scripts. That scheme is hardly as convenient as using dump/restore. You get multiple dump levels (up to 10) by changing one character in your dump script. I wonder how long a complicated script that manipulates backup_history files would be? One big problem with wbak is when there's a tape error writing the sixth of six tapes. You have to start the wbak all over again (please please please correct me if I'm wrong). Dump, on the other hand, checkpoints itself every tape so at worst it has to redo the sixth tape only. Even if wbak is a great tool, and even if it's faster, it would be really nice to have dump and restore also. In a mixed environment such as we have here (Vaxes, Suns, Alliant, Apollos) it would be very convenient to use the same backup scheme on all machines, and to be able to read Apollo backup tapes on a Vax in a pinch, or dump over the network to a high-density 8mm drive attached to a Sun, and so on. So long as I'm in a curmudgeonly mood, can you please let me turn off the registry and just deal with a regular /etc/passwd file? The registry is a powerful tool, but we don't have a thousand Apollos, we only have thirty Apollos and it would be much simpler if we could use the same account manipulation tools that work for us on all our other machines. And ACL's are wonderful flexible things too, I agree, wouldn't it be nice if all Unix implementations had them, now please just show me how to disable them and go back to regular Unix protections. I can hear the Apollo guys saying "No, wait, you don't want to use /etc/passwd, the registry is great, it's got this and this and this feature, it's replicated, it's got "organizations" as well as groups, etc etc etc". Well, sure. But can you please give me the option of using a normal password file. Sun has the right idea - they supply the Yellow Pages but they don't force you to use it. It's probably frustrating for Apollo to develop these advanced tools and then have neanderthals like me clamor for the old unimproved ones, but, well, we were told that SR10 was going to be "4.3, bugs and all", and it sure doesn't seem that way. Where's the "w" command? Where's "adb"? Whaddya mean, there's no assembler? Steve Hayman Workstation Manager, Indiana University P.S. Sorry, I really didn't mean for this to be so cranky sounding. Really, I like using my Apollo on the whole, but we've been having a rough time trying to get SR10 going, and we're on the verge of going back to 9.7. P.P.S. If anybody out there is doing a native 4.3 for the Apollos, have I got a purchase order for you.
giebelhaus@hi-csc.UUCP (Timothy R. Giebelhaus) (12/30/88)
In article <15819@iuvax.cs.indiana.edu> sahayman@iuvax.cs.indiana.edu (Steve Hayman) writes: >Timothy R. Giebelhaus of Apollo writes: >>You can get much more sophisticated by manipulating the backup_history files. >>For example, you can get muliple levels of dumps by saving multiple >>backup_history files. Though this is more complicated, it can be handled >>through scripts. > > >That scheme is hardly as convenient as using dump/restore. You get multiple >dump levels (up to 10) by changing one character in your dump script. >I wonder how long a complicated script that manipulates >backup_history files would be? It would not be bad at all. It would be much more flexable too. For example: you could copy the backup_history file after backup to something like backup_history.88.12.29. Now, if you wanted to backup everything since Dec 29, 1988, you just copy backup_history.88.12.29 to backup_history and start the backup. I think you'll agree that this would not be so very difficult. We can talk about it in more detail on Wednesday. Rbak and Wbak are not perfect, but they do have a lot to offer. >So long as I'm in a curmudgeonly mood, can you please let me turn off >the registry and just deal with a regular /etc/passwd file? The >registry is a powerful tool, but we don't have a thousand Apollos, we >only have thirty Apollos and it would be much simpler if we could use >the same account manipulation tools that work for us on all our other >machines. And ACL's are wonderful flexible things too, I agree, >wouldn't it be nice if all Unix implementations had them, now please >just show me how to disable them and go back to regular Unix >protections. Do you really want seperate /etc/passwd files for each of the 30 nodes you have? Or, do you want one /etc/passwd for all 30 nodes? Even at 30 nodes I would think you would want some replication in case your node with the /etc/passwd goes down. I don't think you want to maintain 30 /etc/passwd files. I assume you would like a passwd interface to the rgyd program. I think I could use the importing password file commands to write a vipw which would let you edit the passwd file. Let me know and I'll see if I can get it finished by Wednesday. >It's probably frustrating for Apollo to develop these advanced >tools and then have neanderthals like me clamor for the old unimproved ones, >but, well, we were told that SR10 was going to be "4.3, bugs and all", >and it sure doesn't seem that way. Where's the "w" command? >Where's "adb"? Whaddya mean, there's no assembler? I don't know about "w" or "adb". I'll try to find out. I know the assembler is around somewhere. I'll find that too. -- UUCP: uunet!hi-csc!giebelhaus UUCP: tim@apollo.uucp ARPA: hi-csc!giebelhaus@umn-cs.arpa ARPA: tim@apollo.com Tim Giebelhaus, Apollo Computer, Regional Software Support Specialist. My comments and opinions have nothing to do with work.