>This question does illustrate why I dislike the IBM suggestion of
individual automounted home directories so much - if your search
includes automounted directories, every single one must be mounted. If
you reckon directory seaches are slow, try HSM recalling filesystems one
at a time...




I dare to disagree a bit. If you look (find, ls, whatever) at a automount 
managed directory, you will see only those subdirectories which currently 
*have* the corresponding file system mounted. You will not detect any other 
subdirectory, because they simply do not exist when the file system is *not* 
mounted.


automount dynamically creates the subdirectory when it is first accessed, then 
mounts the file system. After some inactivity time interval (depending on 
automount configuration), automount unmounts the files system, and deletes the 
subdirectory.


And this makes it ever worse to scan such trees. You need to find a way to 
learn what directories are to be searched, and then start another search for 
each one of them.
Think about that this means to the file level backup process, should you have 
one. That process will again see only what is munted at the time the backup 
runs.

--
Peter Hunkeler

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to