Eli Marmor wrote:
Beni Cherniavsky wrote:
I'm afraid that's too OT (since linux supports pretty much all drives ;-), so here is a related question: what measures (short of RAID) could I use to reduce the risk of disk errors? I'm starting to be annoyed by them, every time it takes long to recover and then not always all is recovered... Are some filesystems (reiser?) less fragile than others? I'd like to protect the /home are (yes, I backup by Unison, replicating important things on several computers but I wonder what other means exist).
Disclaimer: I'm not a hardware expert:
Buy the slowest SCSI that you can find.
Finger rules for better quality:
SCSI vs. IDE: SCSI is better Slow vs. fast: slow is better New model vs. old model: new is better Low capacity vs. high: low is better
However, newer is usually faster. In addition, SCSI is usually faster than IDE (there are many 15,000RPM SCSI disks). Nevertheless, buy a new SCSI model, but slow one.
P.S. There is a danger with journaling file systems: While it's great that after a crash FSCK doesn't have to scan the entire file system, they may miss problems of data integrity in the file system, and the file system may pass months or even years without a real FSCK, which is a bad idea (sometimes you may even have problems in sectors that nobody wrote to, because of hardware problems).
For ext2/3:
man tune2fs and then use it! Even on a large capacity drive, you can set it up to run at night. (This is a bit of advice I got from the list back in 1999!)
Dunno about Reiser and the rest.
There are great advantages for journaling file systems, but don't forget the drawbacks!
================================================================= To unsubscribe, send mail to [EMAIL PROTECTED] with the word "unsubscribe" in the message body, e.g., run the command echo unsubscribe | mail [EMAIL PROTECTED]
