- Recoll vs. Beagle
- Helper Applications
- Installing Recoll
- Setting Up Recoll
- Using Recoll?
Setting Up Recoll
First, open /home/username/.recoll/recoll.conf. Make sure that
- topdirs points to the root of your directory tree, i.e. /home/username.
- Wildcards will tell Recoll what you do not want scanned. Enter the directory names (such as .wine), not the paths (such as /home/username/.wine), at the top of the directory trees you don't want indexed.
That's probably as much as you'll need to customize. Read through the rest of the recoll.conf file to be sure.
Next, run this command:
$ recollindex
Note that the recollindex drive-indexing program is usually invoked as user, not root, as the $ prompt indicates.
The software puts its index into a database program called Xapian. It took several hours to build up the index for a 50 gigabyte filesystem. Update takes about the same length of time. I run recollindex overnight three times a week. If you've got a smaller filesystem to index, it'll take less time.
You can also run recollindex as a continuous daemon, the same way Beagle runs:
$ recollindex -m
For me, the hit on performance in this case was unacceptable. Suddenly, things for which I expected instant response took seconds, and I'm running a decent entry-level workstation (Athlon 64 X2/4200+, SATA HDs). This may change if an improved sata_nv (NVIDIA) driver lets me get more speed out of the drive (I'm only getting 67 megabytes throughput), but at this point I certainly wouldn't run it as daemon. If your SATA HD driver functions better than mine, the daemon might work for you. If you are running a legacy IDE drive, it probably isn't even worth trying.