Greetings,
This message is to inform users of the SGI NIS cluster
that the disk
mounting procedure that we are using has now changed
(hopefully for the
better).
Previously, mounting of disks from one sgi onto another
was
accomplished when the /etc/fstab file (which defines
both the local and
remote disks to be mounted) was read at boottime.
The problem with
this procedure is that sometimes other computers
having disks to be
mounted are not up and running. This causes
the mount to fail, which
as many of you know, can cause trouble (in terms
of not being able to
access one's files). The only way to fix the
failed mount under this
scenario is to bring up the computer whose disk
is to be mounted, and
then manually issue a mount command as root (which
is a pain for me).
The new mounting system is called autofs. This
procedure is nice in
that the mounts are dynamic, where mounting and
unmounting occurs
automatically as needed. Thus, under autofs,
if a user is logged onto
instinct, and then types cd /inept2, this will automatically
initiate a
mount of the /inept2 disk. Note that under
this scenario, all that is
required is that inept is up and running at the
time the user types the
cd command (clearly if inept is not up and running,
the cd command will
still fail).
One detail is that you should be aware of is how
the mounts are made.
Under autofs, all mounts are made under
/hosts/name_of_host_sharing_the_disk/disk_or_directory_to_be_mounted
where name_of_host_sharing_the_disk is the name
of the host sharing the
disk (like instinct, inept, bioc01, nmrfac) and
disk_or_directory_to_be_mounted is the name of the
disk to be mounted
(like instinct1, inept2, bioc011, nmrfac1).
As an example, if user
ahinck logs into the machine inept, you would find
that the result of
typing the pwd
command is
/hosts/instinct/u/people/ahinck
rather than /u/people/ahinck as before. NOTE,
HOWEVER, it is not
necessary to specify the
/hosts/name_of_host_sharing_the_disk when specifying
a disk location.
Thus, if user ahinck was logged onto inept, and
he wanted to change
back to his home directory, he could simply type
cd /u/people/ahinck (cd alone would work as well)
rather than the more cumbersome /hosts/instinct/u/people/ahinck.
The reason for this is that symbolic links have been
added to enable
the simpler designation. Thus, on the computer
inept, a symbolic link
(called /u) has been added at the root level (/)
which points to the
directory
/hosts/instinct/u. Thus, typing cd /u is equivalent
to typing cd
/hosts/instinct/u. The same is true for all
of the other shared disks.
[[ It is possible that I have made a
mistake in typing in one of the
symbolic links, so if you think you find something
that does not appear
to be working as expected, please let me know and
I will check it out
]].
In summary, the new system should be much more reliable
than the old
without any real practical differences compared
to the old system
(owing to the symobolic links, the disk navigation
is identical to that
used under the old system; the only thing different
you should notice
is a different result from the pwd command]]
Please, note that further steps are being taken to
improve the
reliability of the SGI NIS cluster. Soon all
of the SGI NIS computers
(except for nmrfac) will have be supplied with emergency
electrical
power. Since all of the SGIs (except nmrfac)
also have UPSs, this
means that in theory, none of the computers should
ever shutdown in the
case of an electrical outage (either shortterm,
or longterm). I
anticipate that the switchover to emergency power
will be accomplished
without any interruption in service to the cluster;
however if this
changes, I will be sure to let you know.
Andy