The hardware database now being a systemd API rather than a udev one, I guess, shows that the systemd-udev migration is going steadily. Lots of emphasis on exploiting btrfs features and beefing up nspawn.
* When the user presses Ctrl-Alt-Del more than 7x within 2s an
immediate reboot is triggered. This useful if shutdown is
hung and is unable to complete, to expedite the
operation. Note that this kind of reboot will still unmount
all file systems, and hence should not result in fsck being
run on next reboot.
Well that's wonderful. Sticky Keys for Linux. I'm waiting for them to implement a custom recovery shell triggered by the Contra code.
I bet you a nickel that the same situations that would cause SysRq+[S,U,B] to fail to properly sync and mount ro would cause systemd's unwedge-the-world emergency reboot to fail to properly sync and mount ro. Notice that the release note says "[Emergency reboot] should not result in fsck being run on next reboot.".
Yes, but it's hard to remember, and it depends on Sysrq being enabled. I think the Ctrl+Alt+Del thing makes sense. Since everyone is adopting Systemd, maybe at least it will be ubiquitous. :P
It does have one advantage: You don't have to press all keys. If you system is just locked up cause it's swapping like crazy, it usually gets usable after the e (terminate all tasks).
Yeah, doing a simple "sleep 2" and trying to press enter 7 times before it returns, showed me how much I was underestimating my button pressing. It's actually not hard.
There actually is a grain of truth to that, and it's already here. The LFS book has a separate edition specifically for systemd, as listed here: http://linuxfromscratch.org/lfs/read.html
In looking the table of contents of both editions, it seems that the only high level differences are in the exchange of sysvinit and systemd. Both serve the same purpose, but the implementation is different. If LFS made a BSD init version, then I would expect a forked edition as well. At first it feels more political than technical, but since the underlying configuration files are going to be vastly different, it makes complete sense. I don't even understand how this could be construed as an argument for either side.
Yep, because systemd isn't just an init replacement. It comes complete with a whole sh'load of replacements for stuff that doesn't need to be replaced! Because fuck you and your stable, perfectly good utilities! MOAR FEATURES!
20 comments
[ 5.6 ms ] story [ 61.9 ms ] thread... and achieves the same thing
http://en.wikipedia.org/wiki/Magic_SysRq_key
I bet you a nickel that the same situations that would cause SysRq+[S,U,B] to fail to properly sync and mount ro would cause systemd's unwedge-the-world emergency reboot to fail to properly sync and mount ro. Notice that the release note says "[Emergency reboot] should not result in fsck being run on next reboot.".
Silly bet.
I was making the point this is useful and convenient in this situation not that it was magic.
It does have one advantage: You don't have to press all keys. If you system is just locked up cause it's swapping like crazy, it usually gets usable after the e (terminate all tasks).
* Compile kernel
* Compile systemd
* Boot
;)
Their reason for adopting eudev had something to do with the problems of automating the extraction of udev from the larger systemd source, iirc.