Originally posted by jrch2k8
View Post
Announcement
Collapse
No announcement yet.
Lennart Poettering Announces New Project: casync
Collapse
X
-
Originally posted by duby229 View PostI'd still to this day like someone to tell me how in what way is that even a problem? You like Fedora, then use Fedora. Ubuntu? Then use Ubuntu, ffs. I never understood why everyone thinks that every other distribution should be just like theirs, and for crying out loud they're gonna make your distribution just like theirs...... So arrogant.
The whole class of issues just doesn't exist anymore. As long as you use a systemd based distro (which is almost all of them), you get a working .service file. If upstream doesn't provide one, you will find one in one of the other distros, can copy it and it _will_ work.
And it will work well, because suddenly the mentioned 17 distro maintainers don't have to solve the same issues over and over. So nowadays jabberd2 starts in a second on openSUSE instead of sometimes taking several minutes leading to the cluster manager deciding that it must be broken. Why several minutes? Because jabberd2 consists of 4 different daemons that have to be started and the old init script used netstat to check if they are already running and if the DNS server is down (because your systems are recovering from a power outage), it would take forever. Could the script have been fixed? Of course. And I'm sure there were distros with better init scripts. Nowadays it just works everywhere. On every single one of the 300+ distros using systemd. Because systemd is a real service manager and actually knows if a service is running so it doesn't have to use badly written hacks.
So now you know.
- Likes 13
Comment
-
Originally posted by niner View Post
Because writing 17 init scripts for your software for 17 different distribution sucks almost as trying to maintain that 800 lines init script fully of hacks so it will run on 17 different Linux distributions. Or having to write your own init script for some software, because the authors don't care and maybe have not even heard about your favourite Linux distro and thus don't provide an init script for you.
The whole class of issues just doesn't exist anymore. As long as you use a systemd based distro (which is almost all of them), you get a working .service file. If upstream doesn't provide one, you will find one in one of the other distros, can copy it and it _will_ work.
And it will work well, because suddenly the mentioned 17 distro maintainers don't have to solve the same issues over and over. So nowadays jabberd2 starts in a second on openSUSE instead of sometimes taking several minutes leading to the cluster manager deciding that it must be broken. Why several minutes? Because jabberd2 consists of 4 different daemons that have to be started and the old init script used netstat to check if they are already running and if the DNS server is down (because your systems are recovering from a power outage), it would take forever. Could the script have been fixed? Of course. And I'm sure there were distros with better init scripts. Nowadays it just works everywhere. On every single one of the 300+ distros using systemd. Because systemd is a real service manager and actually knows if a service is running so it doesn't have to use badly written hacks.
So now you know.
Comment
-
Originally posted by duby229 View PostLook, we all know full well that LP hates the Unix philosophy. But the real truth is that rsync, dd, truncate and a few others have been doing what LP needs for decades. And if he wasn't so scared of shell scripting he could already have been doing what he wanted.
http://www.dirtcellar.net
- Likes 5
Comment
-
Originally posted by waxhead View PostI am curious. How can you write a example with rsync , dd and truncate replicate some of the features of casync?!
- Likes 3
Comment
-
Originally posted by duby229 View PostLook, we all know full well that LP hates the Unix philosophy. But the real truth is that rsync, dd, truncate and a few others have been doing what LP needs for decades. And if he wasn't so scared of shell scripting he could already have been doing what he wanted.
Also please state what Unix-phylosophy-approved binary tool you plan to use for deltas.
- Likes 3
Comment
-
If rsync is so great, why does the world need zfs send or btrfs send/receive?
Just saying, sometimes it's possible to solve a problem better, and sometimes it's possible to solve a similar, but different, problem. There's too much knee-jerk hate based on a specific name around here. I fully imagine that some of you guys, if you'd never heard someone's name before, would totally find some of the concepts good if given to you via A/B testing.
- Likes 8
Comment
-
Originally posted by duby229 View PostLook, we all know full well that LP hates the Unix philosophy. But the real truth is that rsync, dd, truncate and a few others have been doing what LP needs for decades. And if he wasn't so scared of shell scripting he could already have been doing what he wanted.
- Likes 7
Comment
Comment