home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.ultrix
- Path: sparky!uunet!stanford.edu!CSD-NewsHost.Stanford.EDU!times.stanford.edu!zowie
- From: zowie@daedalus.stanford.edu (Craig "Powderkeg" DeForest)
- Subject: Re: Swap area on Dec5000
- In-Reply-To: white@trantor.cs.unc.edu's message of 18 Dec 92 14:22:06 GMT
- Message-ID: <ZOWIE.92Dec18072225@daedalus.stanford.edu>
- Sender: news@CSD-NewsHost.Stanford.EDU
- Organization: Stanford Center for Space Science and Astrophysics
- References: <16DEC199221211918@csa1.lbl.gov> <18452@borg.cs.unc.edu>
- Date: 18 Dec 92 07:22:25
- Lines: 25
-
- In article <foo> white@trantor.cs.unc.edu (Brian T. White) writes:
- Shaheen R Tonse writes:
- > A colleague tells me that it is useless to just make new b partitions and
- > expect the system to know that they are available for swap area, unless
- > one somehow re-compiles this information into the kernel.
-
- [instructions for adding a new swap space deleted]
-
- One thing to be on the lookout for, is that some non-DEC disks (notably
- our Fujitsu 1.3Gb drive) don't spin-up until the first access request
- after a SCSI reset. So the first access always fails. The DECstation
- doesn't like that -- the first access is usually the fledgling kernel
- trying to figure out where it can swap, and it decides there isn't an
- rz<n> when that first access fails -- basically you just can't swap to
- one of 'em. (unless someone's got a brilliant idea to get around the
- problem without a source license...)
-
-
-
-
-
-
-
- --
- DON'T DRINK SOAP! DILUTE DILUTE! OK!
-