Is it possible to merge slocate databases built on several machines in a local network? Typically, updatedb omits network-mounted directories when building/updating the database, otherwise it could go forever. Still, it might be useful to have a tool to construct a merged database built up from two or more local slocate databases and be able to search files located anywhere in the LAN, just in a single pass, without repeating the search on every machine.

regards, Michal.

i wouldn't look to merge databases, but use each database in turn... you could easily store the databases on a common NFS location and the cycle through them with a simple bash loop. that also let's you add in extra details about what machine it is on, which isn't held within the database. alternatively you could leave all databases untouched and use a framwork of ssh to connection into each machine and run the searches remotely and just give you back the data. this would X number of machines search a singel database, rather than a single machine search X number of databases, so will be faster by a factor of X.

Thanks for the hints. Still, there is a problem.

I arranged copying (in /etc/crontab) individual slocate.db files from several machines to one location on the main server, with appropriate filename change.

When I, however, try to search the databases being logged to the main server, using "slocate -d /path/machine_name.db pattern" it shows nothing. Only when I get superuser privileges, it shows the files matching the pattern.

It seems that "slocate" run by an ordinary user checks for existence of each file matched. Surely it does not find the file on another machine.

Can this be avoided? Or the only way is to make multiple "rsh" or "ssh" calls to every machine which in turn requires the access to those machines without using password.

regards, Michal.