[EMAIL PROTECTED] said: >On Thu, 29 Aug 2002, John Abreau wrote: > >> Rich Payne <[EMAIL PROTECTED]> writes: >> >> > How many clients are you talking about? I would strongly reccomend >> > looking into some sort of network attached storage device. We use Network >> > Appliance machines. At my previous employer we started out with a pair of >> > Suns and hardware raid 5, then went over to the NetApps. They just serve >> > files for a living, and they do it well. >> >> I'd have to agree with this. Once you get past the sticker shock, the >> NetApp appliances are awesome. I found them to be fast and reliable, and >> the snapshot feature is a lifesaver. At any point in the filesystem, you >> can access a read-only ".snapshot" directory to recover older files.
Add to the fact that Netapp lives & breathes file serving and any file server you build ain't going to measure up to the performance. Netapp also focuses on reducing latency which is usually more important to end users. .. stuff about the wonderful snapshots deleted. I wish this was in a linux file system.... >One interesting thing I found out about the NetApps is that the physical >position of each disk isn't important. You can take all the disks out, put >them back in a different order, even in different shelves and when you >power the box back on it will find everything and go on it's way! Heh. When I was at UB, we had 3 of the original Netapps. 1400? They were 486 EISA based with 14 disks each. All maxed out. It was basically a custom PC with the disks ridgidly mounted. No hot sawpping here. We decided to upgrade. Netapp had the 220(?) a Pentium 133 based system and the 420, based on the alpha. We got 1 520 and 2 220s. They used DEC storageworks shelves & drives which we were familiar with from building fileservers on solaris in a previous job. Our VAR said we could get storageworks cases and put our drives in them, saving lots of $$$. So.... And that worked well. We didn't have to reformat, restore from tape, anything like that. The upgrade went very well. A week later, a disk died. To be expected after an upgrade like this. Another died (we had a few hot spares). Another. Our VAR was shipping us drives & working on the problem. Turns out we had 2 problems. The 1st problem was the switch from ridgid mounts to the storageworks shelves. It seems the drives heads were in sync & hit a harmonic motion. They vibrated themselves to death. Seagate had a firmware upgrade & we had to swap out the drives with the VAR. Good thing we had a good relationship with them. A few weeks later, we started having crashes on the 220s. We lost a disk. This time, it was bad motherboards. We eventually got the VAR & Netapp to swap out the 220s for 540s based on the alpha like our other, flawless Netapp. We lost 4-6 disks and had motherboard issues. Through all of this we didn't lose one bit of data. In 4 years with the netapps, we *never* had to restore from tape. -- ------- Tom Buskey _______________________________________________ gnhlug-discuss mailing list [EMAIL PROTECTED] http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss