1 |
Am 26.05.2014 21:57, schrieb Stefan G. Weichinger: |
2 |
> Am 26.05.2014 19:47, schrieb Stefan G. Weichinger: |
3 |
> |
4 |
>> But I somehow think the performance is sub-optimal. |
5 |
> |
6 |
> virt-backup is slow as well (using dd and gzip or pigz in my own patched |
7 |
> version). Yes, that LVM stuff again ... |
8 |
> |
9 |
> I run 6 SAS disks and built hardware raids. |
10 |
> |
11 |
> Should I look into the cache settings there? |
12 |
> |
13 |
> |
14 |
> way too slow ... |
15 |
|
16 |
I think I have some IO-topic going on ... very likely some mismatch of |
17 |
block sizes ... |
18 |
|
19 |
the hw-raid, then LVM, then the snapshot on top of that ... and a |
20 |
filesystem with properties as target ... oh my. |
21 |
|
22 |
Chosing noop as IO-scheduler helps a bit but maybe I have to roll back |
23 |
and rebuild one of the HW-RAID-Arrays with a different blocksize. |
24 |
|
25 |
Stefan |