Hi,
I have a small home LAN consisting of four PC's: one old Pentium III box acting as gateway / connection handler / printer server / file server / scanner server, and then one desktop and two laptops. I recently migrated this LAN from Slackware 10.2 to CentOS 4.3, and everything runs rather fine.
One thing I have some trouble setting up is my Canon Perfection USB scanner for network use, which worked fine with Slackware, because things work a bit differently under CentOS.
The scanner is attached to the "server" box (192.168.1.1) with a minimal install of CentOS, no window manager, no X. I installed sane-backends on this machine and ran scanimage -L as root, which gave this:
[root@babasse ~]# scanimage -L device `epson:libusb:001:002' is a Epson Perfection610 flatbed scanner
I read the various docs on sane-project.org, especially this document:
http://penguin-breeder.org/sane/saned/
My first problem, and apparently one big difference between CentOS and Slackware, is that there is no entry for the scanner in /dev. Now what does the above output of scanimage -L exactly mean? That my scanner device file is /proc/bus/usb/001/002?
[root@babasse 001]# ls -l /proc/bus/usb/001/002 -rw-r--r-- 1 root root 50 jun 5 08:57 /proc/bus/usb/001/002
Now, if that is the case (if not: correct me please), how do I change permissions to that device file? So it doesn't belong to root.root, but to saned.saned, according to the document linked above.
By the way: saned.saned doesn't exist, so I just created this system user like this:
# useradd -d /dev/null -s /bin/false saned
... which gives:
[root@babasse 001]# cat /etc/passwd | grep saned saned:x:500:500::/dev/null:/bin/false
(Did I define this system user in an orthodox way? Or is there a better way to do this?)
Now when I su to that user and try to run scanimage -L as saned, I get a "No scanners were found" error, which is most likely a permission problem on the device. Now the big question: how do I fix this? I googled about "libusb hotplug scanner device permissions" and some permutations. The problem is not the lack of information on the subject, but the diverging wealth of it.
Any suggestions?
Niki Kovacs