Figures…
…a day after I just updated my desktop at home to 3.19. About time I updated the webserver though.
Linux 4.0 is upon us. Linux Lord Linus Torvalds made it official with a typically brief post and slightly longer explanation to the Linux kernel mailing list. The new number isn't a sign of a major upgrade. As we've chronicled, Torvalds thinks that it looks a bit silly when version numbers go beyond x.19. He therefore decided …
3.19? That's soooo last month(ish).
The non disruptive patching thing is pretty interesting and rather understated in the article. Imagine not having to reboot after updating your kernel! You might like to compare that with even a basic app install on some other OSs which require a reboot.
This might not be such a big deal per se for a home laptop with SSD but the init. on some server kit takes ages especially if Fibre Channel across multiple paths to multiple targets is involved.
It's not finished yet but soon you will be able to do all updates without a reboot but a service or two restart will be necessary. Incidentally, under Linux "lsof | grep 'DEL.*lib' | cut -f 1 -d ' ' | sort -u" will give you hints as to what might need restarting after an updating session. Many distros already take care of that already.
Cheers
Jon
I agree, the live patching is really (to me) a big deal. while I haven't had to reboot server (centos) often for patches its still a pain especially when kvm having issues.
this is just one more thing to make it easier and better.
wish ms would do similar. hate updating the exchange server.
"Torvalds thinks that it looks a bit silly when version numbers go beyond x.19"
Er, no, he doesn't. Nowhere in the linked article, or the LKML post or anywhere else has he said that. Have you forgotten the days of 0.99pl13? He decided to go to 4.x because people seemed to say they preferred it.
This post has been deleted by its author
Before people get too excited over the live patching feature, I should point out this kernel just has the first stage of providing live patching support. See this LWN article for more details: lwn.net/Articles/634649/.
The funny thing is what now will think all those Linux "hardcores" who has been relying on a non yet existing feature for all the past years (unless you explicitly used one of the two non standard systems), swearing you never, never needed to reboot a systema after a kernel patch... it's good those discussion now will surface a lot of bad practices that left system at risks until a reboot was needed for some other reason.
Just because I was sure I would have hit hard all the lame sysadmins that really believe Linux never needs a reboot... and now are faced with the horrible truth they mismanaged their servers for years believing in pink unicorns... whoever had just a pale idea of how a kernel works would have known it, but the army of lusers trained to believe Linux is the OS coded by god himself blindly believed what other lusers without a clue about what a kernel is believed without actually checking...
And now after the OpenSSL/Bash debacle - open source doesn't mean bugs are sorted out immediately by the millions of code reviewers - they has to accept Linux needs reboots after patching, even if it doesn't tell you... waiting for the deniers to pretend they never said it and feel free to downvote, truth hurts...
PS: I do use Linux, but I don't worship any software, it doesn't make feel me better or superior if I use one or another, but I really hate lies and nerdy worshipper.
That criteria of users isn't just isolated to Linux reboots after kernel updates.
For some absurd reason it seems people would rather believe garbage posted and then resposted on blogs as gospel than actually thinking, going to the reference docs or product source. After all the number 1 link in a google search, which returns a link to the top of the blog posting pyramid 1/2 baked article by someone with an axe to grind is obviously more accurate than reference documents or provable results.
A very bad side effect of blogging is people with no clue but verbal diarrhea can become the accepted reality for best practice by people who the repost without questioning or understanding the drivel they are supporting