| 1 |
On Thu, 05 Mar 2015 03:16:55 -0600, Dale wrote: |
| 2 |
|
| 3 |
> > Of course it wasn't. Warnings about /var not being writeable are not |
| 4 |
> > going to be written to /var. |
| 5 |
|
| 6 |
> Yea, it won't catch everything. This is sort of designed for that point |
| 7 |
> where one log stops and the other hasn't started yet. This is usually |
| 8 |
> where dmesg stops and syslog and friends hasn't yet started. Of course, |
| 9 |
> if /var isn't mounted, well, it has no where to go. |
| 10 |
> |
| 11 |
> Isn't those init thingys supposed to fix this sort of thing tho? Isn't |
| 12 |
> it supposed to store it in memory until /var is mounted and then dump |
| 13 |
> it? |
| 14 |
|
| 15 |
The combination of dracut and systemd means it all goes in the journal. |
| 16 |
The journal is kept in memory and then flushed to disk later in the boot |
| 17 |
process. |
| 18 |
|
| 19 |
journalctl is rather neat, being able to search both dmesg and system |
| 20 |
logs in one go, and apply various search criteria, is far more flexible |
| 21 |
than the alternatives. |
| 22 |
|
| 23 |
|
| 24 |
-- |
| 25 |
Neil Bothwick |
| 26 |
|
| 27 |
WWW: World Wide Wait |
Archives