| 1 |
On Sonntag 07 Februar 2010, Mark Knecht wrote: |
| 2 |
> On Sun, Feb 7, 2010 at 9:30 AM, Alexander <b3nder@××××××.ru> wrote: |
| 3 |
> > On Sunday 07 February 2010 19:27:46 Mark Knecht wrote: |
| 4 |
> >> Every time there is an apparent delay I just see the hard drive |
| 5 |
> >> light turned on solid. That said as far as I know if I wait for things |
| 6 |
> >> to complete the data is there but I haven't tested it extensively. |
| 7 |
> >> |
| 8 |
> >> Is this a bad drive or am I somehow using it incorrectly? |
| 9 |
> > |
| 10 |
> > Is there any related info in dmesg? |
| 11 |
> |
| 12 |
> No, nothing in dmesg at all. |
| 13 |
> |
| 14 |
> Here are two tests this morning. The first is to the 1T drive, the |
| 15 |
> second is to a 120GB drive I'm currently using as a system drive until |
| 16 |
> I work this out: |
| 17 |
> |
| 18 |
> gandalf TestMount # time tar xjf /mnt/TestMount/portage-latest.tar.bz2 |
| 19 |
> -C /mnt/TestMount/usr |
| 20 |
> |
| 21 |
> real 8m13.077s |
| 22 |
> user 0m8.184s |
| 23 |
> sys 0m2.561s |
| 24 |
> gandalf TestMount # |
| 25 |
> |
| 26 |
> |
| 27 |
> mark@gandalf ~ $ time tar xjf /mnt/TestMount/portage-latest.tar.bz2 -C |
| 28 |
> /home/mark/Test_usr/ |
| 29 |
> |
| 30 |
> real 0m39.213s |
| 31 |
> user 0m8.243s |
| 32 |
> sys 0m2.135s |
| 33 |
> mark@gandalf ~ $ |
| 34 |
> |
| 35 |
> 8 minutes vs 39 seconds! |
| 36 |
> |
| 37 |
> The amount of data written appears to be the same: |
| 38 |
> |
| 39 |
> gandalf ~ # du -shc /mnt/TestMount/usr/ |
| 40 |
> 583M /mnt/TestMount/usr/ |
| 41 |
> 583M total |
| 42 |
> gandalf ~ # |
| 43 |
> |
| 44 |
> |
| 45 |
> mark@gandalf ~ $ du -shc /home/mark/Test_usr/ |
| 46 |
> 583M /home/mark/Test_usr/ |
| 47 |
> 583M total |
| 48 |
> mark@gandalf ~ $ |
| 49 |
> |
| 50 |
> |
| 51 |
> I did some reading at the WD site and it seems this drive does use the |
| 52 |
> 4K sector size. The way it's done is the addressing on cable is still |
| 53 |
> 512 byte 'user sectors', but they are packed into 4K physical sectors |
| 54 |
> and internal hardware does the mapping. |
| 55 |
> |
| 56 |
> I suspect the performance issue is figuring out how to get the file |
| 57 |
> system to keep things on 4K boundaries. I assume that's what the 4K |
| 58 |
> block size is for when building the file system but I need to go find |
| 59 |
> out more about that. I did not select it specifically. Maybe I need |
| 60 |
> to. |
| 61 |
> |
| 62 |
> Thanks, |
| 63 |
> Mark |
| 64 |
|
| 65 |
no. 4k block size is the default for linux filesystems. But you might have |
| 66 |
'misaligned' the partitions. There is a lot of text to read about |
| 67 |
'eraseblocks' on ssds and how important it is to align the partitions. You |
| 68 |
might want to read up on that to learn how to align partitions. |
Archives