1 |
Volker Armin Hemmann <volkerarmin <at> googlemail.com> writes: |
2 |
|
3 |
|
4 |
> Am 13.11.2014 um 01:01 schrieb Adam Carter: |
5 |
|
6 |
> "Backblaze's analysis of nearly 40,000 drives |
7 |
|
8 |
Always suspect a vendor's purpose for fingering others. It |
9 |
may be valid, but often tainted. |
10 |
Were SSD and other storage devices included? Where operating |
11 |
systems were the drives used on? A myriad of parameters |
12 |
might be insightful....ymmv. |
13 |
|
14 |
|
15 |
> SMART 5 - Reallocated_Sector_Count. |
16 |
> SMART 187 - Reported_Uncorrectable_Errors. |
17 |
> SMART 188 - Command_Timeout. |
18 |
> SMART 197 - Current_Pending_Sector_Count. |
19 |
> SMART 198 - Offline_Uncorrectable" |
20 |
|
21 |
> |
22 |
http://www.computerworld.com/article/2846009/the-5-smart-stats-that-actually-predict-hard-drive-failure.html |
23 |
|
24 |
> everybody with half a brain would figure that one out themselves. |
25 |
|
26 |
|
27 |
Volker is onto something here. Another consideration is which of these |
28 |
(maybe all 5?) can be tracked or have a module developed and included |
29 |
into your favorite NMS, like nagios or jffnms. Collect up the data |
30 |
trend it yourself and auto_generate graphs and or reports. So what |
31 |
is really important is which ones can be included via snmp or whatever |
32 |
protocol into your favorite NMS? |
33 |
|
34 |
Something that relies of something from this vendor, that is not |
35 |
opensource, is worthless noise, imho. |
36 |
|
37 |
|
38 |
hth, |
39 |
James |