John Drescher wrote: > > I know that RAID5 is very appealing 'cause you end with 'a lot of > usable disk space' but, I always discourage to use it. > > Each chunk of data to be written to disk is divided in Numer of > physical RAID disks writes - 1 plus another write to store CRC on > to the remaining disk ( the CRC chunk is written in a round-robin > model using all RAID disks ). For write changed data, the RAID > logic must locate where the data was, rewrite it and calculate the > changes to the parity to reflect the changes > > Another important point is the time spend on rebuild a previously > failed disk ( data is not copied from antoher disk, is regenerated > using the data in CRC chunks ) > > And another tip, If fails 2 disks, good by data, must restore from > backup > > > How about raid 6? With raid 6 you can loose 2 disks and you loose > nothing. I have around 10TB mostly on raid 6 using 250GB and 330 GB > SATA drives and they work great. I have had to replace a disk from > time to time but I have never had 2 go bad at once. > > Always try to use any more 'friendly' RAID as RAID1 / RAID10 > > But then you throw away 1/2 the space.
So the question becomes: What is your objective? To build a cheap system? Or to build a robust and reliable system that you can depend on? I realize that this is not a black or white decision, it is somewhat of a continuum. We buy Seagate Cheetah UltraSCSI drives and have never had an issue with them. While they aren't the cheapest components, they have been continually coming down in price. At present, I'm typically not using Raid. Instead, I'm looking for redundancy in backups. I have a backup server with a significant amount of disk space, and I also have a tape library with 6.4TB capacity. So, I run things to disk and then to tape. And I keep tapes for 6 weeks with additional end of semester archives kept for a year or two. A backup system is an important system and is supposed to provide data security. The sysadmin who loses data is dead meat. Even on my home network, my daughter was about to "kill" me when it seemed that I might have lost some crucial backups. When I had an external array a few years ago, it had two scsi controllers and 12 drives (a Sun D1000). Since I had two controllers, I connected it to two separate scsi buses on the server and configured it as raid10. That gave me fully duplicate data and was really fast -- data written simultaneously out two scsi buses with no crc computation. If a drive failed, I could just replace it and copy the corresponding drive. If two drives failed, the only way I would lose data would be if they were the corresponding drives of a pair. In principle, I could lose 6 drives without losing data, although the odds of that are beyond astronomical. I never had even one drive failure. But, assuming I had 2 drives fail, the odds would be 1/11 that the loss of the second drive would cause me to lose data. Sure, it used up half the drives for redundancy, but that's sort of the point. Raid5 would be 20% (of total capacity) used for redundancy. Raid6 would be 33% redundancy. Raid10 is 50% redundancy. And Raid10 is faster and simpler. So, it's a choice you get to make and/or try to sell to the person who makes purchasing decisions. --------------- Chris Hoogendyk - O__ ---- Systems Administrator c/ /'_ --- Biology & Geology Departments (*) \(*) -- 140 Morrill Science Center ~~~~~~~~~~ - University of Massachusetts, Amherst <[EMAIL PROTECTED]> --------------- Erdös 4 ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
