Topic: Most reliable hard drives (having lots of trouble lately w/seagate 500gb sata)

[dilligaf]

we have used most brands over the years, the last 3 years seagate being the most reliable.
however, we have had a real bad run w/ the 500 gb this last while.
I am happy to pay more for the best, but what is that right now?
HD Failures suck.

[CharlieBrady]

... but what is that right now?

I'm quite sure that you will get better answers to that question elsewhere. Your question is entirely unrelated to SME server software.

[dilligaf]

I'm quite sure that you will get better answers to that question elsewhere. Your question is entirely unrelated to SME server software...

Yeah, that's why I put it in:
General Discussion
Discussions about contribs.org, Linux in general, and "other topics" that don't belong in the other forums.

[dilligaf]

For any that are curious, these two comments from another forum:

-I had a lot of failures on Seagate Sata ES both 320 and 500 Gig HD's. We find Hitachi most reliable.

-we have a couple of servers running RAID-5 with Barracuda ES (server version) and I had 7 broken harddisks on a total of 10 drives. (age 1-2 years) Other servers are running hitachi desktop harddrives and for 4 years no problem.

I am not saying what you should or shouldn't use, I personally used to have excellent luck w/Seagate.
However have had at least 6 failures in last month on 500GB an up (one TB drive as well)

[calisun]

I just had a Hitachi 750GIG SATA drive fail after less than a year. Thank god I set up a RAID, so server kept on humming with one drive dead.
I am sitting here impatiently right now waiting for UPS delivery with my replacement drive.
After this drive failure incident, I will never run a server with a single drive, from now on, it's RAID all the time.

[p-jones]

After this drive failure incident, I will never run a server with a single drive, from now on, it's RAID all the time.

Calisun
Consider the possibility that with RAID 1, if a drive starts to go bad and corrupt the data, that corruption can be reflected in the mirror.

IMHO, it is best only for sudden death of the drive.

[Foa]

Memorycards is up to 32 GIG
Soon no diskdrives to be used?

[calisun]

Calisun
Consider the possibility that with RAID 1, if a drive starts to go bad and corrupt the data, that corruption can be reflected in the mirror.

I would rather have some good data on a good drive than no data on a dead drive

[David Harper]

This illustrates the need for a dual backup strategy: RAID for emergencies plus some kind of external backup (probably tape so there can be regular backups sent off-site).

[calisun]

This illustrates the need for a dual backup strategy: RAID for emergencies plus some kind of external backup (probably tape so there can be regular backups sent off-site).

I do backup besides RAID, I do regular backups from server-manager (backup to desktop)

My issue is that my server is in a co-location facility which is almost an hour drive from my house. So driving there, replacing drive and restoring from backup would take several hours or maybe a day. By having RAID, my server keeps on humming even after my drive failure, that gives me time to replace failed drive without panic.

[Stefano]

just to clarify: raid is not a backup strategy, simply because it's not designed for.

so, for data security, we need raidX (X>0) AND backup

all IMHO
ciao
Stefano

[David Harper]

Indeed, that's much more accurate (and eloquent) than my original comment

[Mjohnson]

just to clarify: raid is not a backup strategy

Bingo.  I'd like to see you use RAID to recover your clients accidentally deleted document or the folder deleted by the intern which contained 2 months worth of proposals. 

Needs for backups do not always involve a hardware failure.  Users are the most dangerous element to a network!!!