zessel@incas.informatik.uni-kl.de (Holger Zessel AG Nehmer) (05/25/90)
bush@evax.arl.utexas.edu (Joe Bush) writes: |> We are running a Ultrix 3.1 on a uVAX 3900 with 32Mb |>of core, by the way... |> The general consensus seems to be that more swap space is |>called for. I am following those suggestions and configuring the |>additional swap space. This seemed like the logical first step since [...] |> I modified the following line to my /sys/conf/SYSTEM file and |>rebuilt the kernel: |> config vmunix ... swap on ra0b and ra1b and ra2b ... |> I added the following two lines to my /etc/fstab: |> /dev/ra1b::sw:0:0:ufs:: |> /dev/ra2b::sw:0:0:ufs:: |> |> Now heres the problem: I then attempted to execute the |> # swapon -a |> Adding /dev/ra1b as swap device |> /dev/ra1b: No such device |> Adding /dev/ra2b as swap device |> /dev/ra2b: No such device hmmm. May be you should try the line *I* inserted for the same purpose: /dev/ra1b::sw::::: May be the ufs type irritates the kernel. I don't know how this line is parsed. |> - Joe Holger ---------------------------------------------------------------- Holger Zessel, uucp: ...!uunet!unido!uklirb!incas!zessel or zessel@informatik.uni-kl.de or zessel%informatik.uni-kl.de@relay.cs.net (from usa) Fachbereich Informatik, SFB124-D1 Universitaet Kaiserslautern, West-Germany
grr@cbmvax.commodore.com (George Robbins) (05/25/90)
In article <1990May24.220458.13891@evax.arl.utexas.edu> bush@evax.arl.utexas.edu (Joe Bush) writes: > > My problem now is getting the system to recognize the > additional "interleaved" swap. Thought maybe you could help me. > I modified the following line to my /sys/conf/SYSTEM file and > rebuilt the kernel: You did copy the new kernel over to /vmunix and reboot, eh? Everything you mentioned looked reasonable... -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing: domain: grr@cbmvax.commodore.com Commodore, Engineering Department phone: 215-431-9349 (only by moonlite)
stefan@wheaton.UUCP (Stefan Brandle ) (05/26/90)
In article <1990May24.220458.13891@evax.arl.utexas.edu> bush@evax.arl.utexas.edu (Joe Bush) writes: > Now heres the problem: I then attempted to execute the >"swapon -a" command (or /etc/rc does) to make the space available to >the system, but getting the following error: > > # swapon -a > Adding /dev/ra1b as swap device > /dev/ra1b: No such device > Adding /dev/ra2b as swap device > /dev/ra2b: No such device Certain things somehow lock devices so that once it is in use, any further attempt to access that device gives the above message. I had the same message when playing around with additional swap space. I've also seen it when playing with the tape unit (/dev/rmt0h & Co.). My guess is that something in the kernel already has your swap devices open in an exclusive mode and any further references give the somewhat misleading message about `no such device' when it might be better to spit out `device locked' or something like that. On the other hand, I'm speaking from intuition, not knowledge, so maybe somebody should correct what I just said. -stefan -- ---------------------------------------------- MA Bell: (708) 260-4110 --------- Stefan Brandle UUCP: ...!{obdient,uunet!tellab5}!wheaton!stefan Wheaton College or stefan@wheaton.UUCP Wheaton, IL 60187 "But I never claimed to be sane!"
schoch@trident.arc.nasa.gov (Steve Schoch) (05/30/90)
In article <2068@wheaton.UUCP> stefan@wheaton.UUCP (Stefan Brandle ) writes: > I had the same message >when playing around with additional swap space. > My guess is that something in the >kernel already has your swap devices open in an exclusive mode and any further >references give the somewhat misleading message about `no such device' when it >might be better to spit out `device locked' or something like that. I believe what happens is unlike some other versions of BSD Unix, when you configure more than one swap devices in the conf file, the Ultrix kernel turns on swap on those devices when the kernel boots, and doesn't wait for swapon. What leads me to this conclusion is we had an 3.0 kernel compiled on a system with two disks and we swapped on both of them. When we copied the same system with only one disk, the kernel complained something about "can't open device (9,1), retrying...." and I guess it timed out after a while. This was before it started /etc/init. Steve
grr@cbmvax.commodore.com (George Robbins) (05/30/90)
In article <50480@ames.arc.nasa.gov> schoch@trident.arc.nasa.gov (Steve Schoch) writes: > In article <2068@wheaton.UUCP> stefan@wheaton.UUCP (Stefan Brandle ) writes: >> I had the same message >>when playing around with additional swap space. >> My guess is that something in the >>kernel already has your swap devices open in an exclusive mode and any further >>references give the somewhat misleading message about `no such device' when it >>might be better to spit out `device locked' or something like that. > > I believe what happens is unlike some other versions of BSD Unix, when > you configure more than one swap devices in the conf file, the Ultrix > kernel turns on swap on those devices when the kernel boots, and doesn't > wait for swapon. No, I don't think it uses them, however it apparently does try to open and size them or some such. The fact that they must be on-line if declared is mentioned in the Ultrix 3.0 [i think] release notes. Beyond that, all is supposition. -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing: domain: grr@cbmvax.commodore.com Commodore, Engineering Department phone: 215-431-9349 (only by moonlite)
wyatt@cfa250.harvard.edu (Bill Wyatt) (05/31/90)
> I believe what happens is unlike some other versions of BSD Unix, when > you configure more than one swap devices in the conf file, the Ultrix > kernel turns on swap on those devices when the kernel boots, and doesn't > wait for swapon. Unless it's changed recently (I haven't checked since the 3.x upgrade), the answer is no, the kernel doesn't use additional swap space until swapon is called (it does use disk 0). We have a uVax with *lots* of disks & partitions, and discovered we couldn't automatically boot after a crash because `fsck -p' ran out of core when checking several filesystems in parallel (about 8). We changed fstab to limit fsck to about 4 or 5 and it's ok, but of course slower. > What leads me to this conclusion is we had an 3.0 kernel compiled on > a system with two disks and we swapped on both of them. When we copied > the same system with only one disk, the kernel complained something about > "can't open device (9,1), retrying...." and I guess it timed out after > a while. This was before it started /etc/init. Some people said that the kernel will hang if you configure a swap partition that isn't there. If this was true in 3.0 (we upgraded from 2.2 to 3.1 directly), it isn't in 3.1. We get messages such as the above, but the kernel gives up after one retry (~10 sec.) and continues booting normally. Bill Wyatt, Smithsonian Astrophysical Observatory (Cambridge, MA, USA) UUCP : {husc6,cmcl2,mit-eddie}!harvard!cfa!wyatt Internet: wyatt@cfa.harvard.edu SPAN: cfa::wyatt BITNET: wyatt@cfa -- Bill Wyatt, Smithsonian Astrophysical Observatory (Cambridge, MA, USA) UUCP : {husc6,cmcl2,mit-eddie}!harvard!cfa!wyatt Internet: wyatt@cfa.harvard.edu SPAN: cfa::wyatt BITNET: wyatt@cfa