1 |
Em 16/06/2014 19:52, "Neil Bothwick" <neil@××××××××××.uk> escreveu: |
2 |
> |
3 |
> On Mon, 16 Jun 2014 16:58:07 -0300, Francisco Ares wrote: |
4 |
> |
5 |
> > Regarding umounting, it umounts the drive after it is removed, which |
6 |
> > looks strange (to say the least) to me. |
7 |
> |
8 |
> Udev doesn't know you want to remove the drive until you've pulled it. |
9 |
> Apparently, a future release of systemd will be able to predict when you |
10 |
> want to remove the drive as it will control everything you do. |
11 |
> |
12 |
|
13 |
:-) well, by 'unount' I mean getting rid of the mounted tree and mount |
14 |
point. |
15 |
|
16 |
> > Did I miss something? Or, in other words, is there a way of forcing |
17 |
> > operations on a usb flash media to always flush buffers and/or caches as |
18 |
> > fast as possible? |
19 |
> |
20 |
> Mount it with the sync option. On the other hand, this increases the |
21 |
> number of writes to the drive, quickly killing performance and slowly |
22 |
> killing the drive. |
23 |
> |
24 |
> |
25 |
|
26 |
Thanks for this, I will use it. This system is an embedded one, the usb |
27 |
stick is for eventual data base backups. I'll remember to do as few writes |
28 |
as possible to it. |
29 |
|
30 |
> -- |
31 |
> Neil Bothwick |
32 |
> |
33 |
> We all know what comes after 'X', said Tom, wisely. |
34 |
|
35 |
Thanks again, |
36 |
Francisco |