[mythtv-users] RAID5 ATA IDE hardware card recommendations for mythbackend machine

Zachary Bedell zaclist at adirondack.net
Thu Dec 16 18:19:15 UTC 2004

Hash: SHA1

On Dec 13, 2004, at 2:36 PM, Russ Dill wrote:
>> I didn't have the money, so I ended up getting a cheap HighPoint card
>> and using Linux software RAID to replace my FastTrack.  The FastTrack
>> is definitely faster than softRAID in terms of throughput, but I'd
>> rather have safer, slower data personally.
> As far as just adding ide ports, would you recommend highpoint?

Not a single complaint with the thing thus far.  It's been in service 
since late September, and I use the array CONSTANTLY, all day long.  It 
holds my mail spool, GB's of media, MP3's, the lot in other words.  The 
one thing I did find is that Linux softRAID responds very poorly to 
power loss.  No data was lost; but after the second time I had to wait 
2+ hours for the thing to resynchronize, I decided I'd better splurge 
and get a UPS.  No problems since...

Of course, if you're going to use the HighPoint for RAID, you need to 
ignore the fact that its four channels can each handle a slave & 
primary.  Promise won't let you put on slaves, neither will 3ware to my 
knowledge.  I'm sure you've read/heard this elsewhere, but resist the 
urge to add more than four drives to the thing.  If a master drive 
fails, it will almost certainly take the slave out too.  That would 
turn your recoverable one drive RAID-5 failure into a fatal two drive 
failure, and woe to you....  'Sides...  These things are cheap enough, 
if you want 5+ drives, just buy an extra controller!  There's stable 
Linux driver support (including multiple cards), so they're a pretty 
stable budget RAID solution.

Best regards,
Zac Bedell

Parallel lines never meet, unless you bend one or both of them.

- ------------------------------------------------------
Brought to you by MacOS, running on host Aramis
Running for:  4 days, 19 hours, 38 minutes

Version: GnuPG v1.2.3 (Darwin)


More information about the mythtv-users mailing list