cks@white.toronto.edu (Chris Siebenmann) (09/26/89)
Well, I found the cause of the problem we were having with "mount -a
-t nfs" getting a SIGSEGV. It turns out that the DS3100 mount is
somewhat stricter (in a sense) about malformed /etc/fstab lines
(although I wish it gave an error message!). We like to automatically
generate various files here (with 9 clients and 2 servers plus the
outside world it becomes a lifesaver when you shuffle NFS mounts
around), so I had an /etc/fstab with the lines
# This is a machine-generated file. Please, fix the master
# template (~cks/adm/templates/fstab.template) and regenerate
# it instead of hand-editing.
at the start (minus the tabs), on the idea/principle that a) many
things take "#" as a commend and b) the local Suns do the same thing
and it works for them. This works on Vaxen, but not on DS3100's; they
seem to dereference a null pointer during a strcmp(). Since it happens
only with a -t specified, I assume it's checking the file system type.
My solution was to tack six ":"'s on, to make the comment lines valid
(albeit unusual) /etc/fstab entries. Then everything works fine. Maybe
we can persuade DEC to recognize comments in /etc/fstab in the next
release...
--
"I shall clasp my hands together and bow to the corners of the world."
Number Ten Ox, "Bridge of Birds"
Chris Siebenmann ...!utgpu!{ncrcan,ontmoh!moore}!ziebmef!cks
cks@white.toronto.edu or ...!utgpu!{,csri!}cksgrr@cbmvax.UUCP (George Robbins) (09/26/89)
In article <89Sep26.001054edt.30815@snow.white.toronto.edu> cks@white.toronto.edu (Chris Siebenmann) writes: > > Well, I found the cause of the problem we were having with "mount -a > -t nfs" getting a SIGSEGV. It turns out that the DS3100 mount is > somewhat stricter (in a sense) about malformed /etc/fstab lines > (although I wish it gave an error message!). There's actually a warning about this feature in the release notes. It sure would be nice if all the ::: like command files, including /etc/passwd understood # comments, but that's another story. On nice thing that crept in somewhere around 3.0 is you can put something like xx in the rw/ro/sw field and have most of the programs cooperate - mount -a doesn't mess with it, but an explict mount by file system name works as does df. Too bad it's not documented or 100% consistant across programs that can access /etc/fstab... -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@uunet.uu.net Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
grr@cbmvax.UUCP (George Robbins) (09/27/89)
In article <8003@cbmvax.UUCP> grr@cbmvax.UUCP (George Robbins) writes: > On nice thing that crept in somewhere around 3.0 is you can put something > like xx in the rw/ro/sw field and have most of the programs cooperate - > mount -a doesn't mess with it, but an explict mount by file system name > works as does df. Too bad it's not documented or 100% consistant across > programs that can access /etc/fstab... Oops, it is documented - maybe I need to do a global diff on man pages and find out what's really new or different... 8-) -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@uunet.uu.net Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)