1 |
Yo Rich! |
2 |
|
3 |
On Fri, 21 Jun 2013 13:57:20 -0400 |
4 |
Rich Freeman <rich0@g.o> wrote: |
5 |
|
6 |
> In theory it could be done that way, but every |
7 |
> raid1 implementation I've heard of makes writes to all drives |
8 |
> (obviously), but reads from only a single drive (assuming it is |
9 |
> correct). That means that read latency is greatly reduced since they |
10 |
> can be split across two drives which effectively means two heads per |
11 |
> "platter." |
12 |
|
13 |
Yes, that is what I see in practice. A much reduced average read time. |
14 |
And if you are really pressed for speed, add more stripes and get even |
15 |
more speed. |
16 |
|
17 |
> Also, raid1 typically does not include checksumming, so if |
18 |
> there is a discrepancy between the drives there is no way to know |
19 |
> which one is right. |
20 |
|
21 |
Uh, not exactly correct. Remember each HDD has ECC for each sector. If |
22 |
there is a read error the HDD will detecct the bad ECC and report the |
23 |
error to the RAID1 hardware/software. Then RAID1 is smart enough to try |
24 |
to read from the 2nd drive. |
25 |
|
26 |
> With raid5 at least you can always correct |
27 |
> discrepancies if you have all the disks |
28 |
|
29 |
Not really. If 2 disks fail in an n+1 RAID5 you are out of luck. |
30 |
Not as uncommon occurance as one might think. |
31 |
|
32 |
> (though as Duncan pointed out |
33 |
> in practice this only happens if you do an explicit scrub on mdadm). |
34 |
|
35 |
Which you should be doing at least weekly. Otherwise you only find out |
36 |
your disks have failed when you try to do a full copy or backup, and then |
37 |
you likely have multiple failures and you are out of luck. |
38 |
|
39 |
RGDS |
40 |
GARY |
41 |
--------------------------------------------------------------------------- |
42 |
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97701 |
43 |
gem@××××××.com Tel:+1(541)382-8588 |