[00:32] <tomreyn> jancoow: on 16.04, systemctl is part of systemd. you should defintely have it available as root. if not, it suggests that your upgrade did not work out well.
[00:34] <tomreyn> there are counter measures against hdd failures. one of them is RAID with a mirroring confiuguration - that's pretty easy if you can afford another disk. the other is to do backups, and do them properly and reliably, which you need to do either way.
[00:56] <jancoow> I replaced my serve casing
[00:56] <jancoow> i've 7hhd's in greyhole configuration
[00:57] <jancoow> the only not redunant disk, the os drive, broke during the move...
[00:57] <jancoow> the sata connectors littarly broke from the pcb and now th drive isn't spinning anymore
[00:57] <jancoow> Last backup is from 2 months ago..
[01:17] <JanC> there shouldn't really be anything important on the OS drive, right?
[01:18] <sarnold> drab: the best documentation of the difference of hash and bitmap might be the source code
[01:18] <sarnold> drab: chances are you probably want one of the hash versions
[06:18] <ChmEarl> is it possible to run the early preseed script and modify the /etc/mke2fs.conf?
[06:19] <ChmEarl> or is there another way to set ext4 options in preseed?
[08:36] <lordievader> Good morning.
[09:19] <rbux> mornin
[13:21] <gordonjcp> hi
[13:21] <gordonjcp> is there a way to get normal network interface names back in 16.04?
[13:22] <gordonjcp> at them moment they're all stuff like enp39asdfasdfblahblahblah and they change on every boot
[13:22] <lordievader> gordonjcp: https://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/
[13:23] <lordievader> They shouldn't change every boot...
[13:23] <gordonjcp> okay, why are they called "predictable" when they're essentially random
[13:23] <lordievader> They are named based on where the bios says they are.
[13:23] <gordonjcp> o_O
[13:23] <gordonjcp> the bios has no idea what a network interface is
[13:24] <lordievader> It does know where the device is.
[13:24] <lordievader> In short, read the link I gave you.
[13:24] <gordonjcp> I'm reading it
[13:25] <gordonjcp> it seems like a pretty pointless change, which only makes things more difficult for users
[13:26] <lordievader> In my opinion it makes more sense than the old way.
[13:26] <gordonjcp> randomly assigning names to interfaces, based on when systemd gets around to looking at them?
[13:27] <lordievader> The point is that it is not random.
[13:27] <gordonjcp> okay, but it's not stable
[13:27] <lordievader> It is.
[13:27] <gordonjcp> it changes on every boot
[13:27] <lordievader> Those ethX names werent. "Oh hey a new interface in the same location, lets give it a new name.
[13:27] <lordievader> "
[13:28] <lordievader> Anyhow, these are just opinions.
[13:28] <gordonjcp> is there a way to get rid of systemd in 16.04?
[13:28] <gordonjcp> or if there isn't, is there a way to force it to work *exactly* like sysvinit
[13:29] <gordonjcp> right now it doesn't seem to run init scripts in any predictable order
[13:29] <gordonjcp> "oh hey, I can't see one of the network interfaces.  Never mind, I'll just start up dnsmasq on it, then I'll bring up the xl2tpd tunnel, oh hell everything on fire now"
[13:33] <gordonjcp> I guess that's a no
[13:33] <gordonjcp> so Ubuntu Server isn't actually intended for use on servers, then?
[13:34] <gordonjcp> or is there something fundamental I'm not getting about this
[14:46] <drab> sarnold: yeah I got some answers in #netfilter, turns out it's mostly a choice based on types and features
[14:46] <drab> looking closer it indeed makes sense for example bitmap:ip expects a range to be specified upon creation, something I hadn't noticed at first
[14:48] <patdk-lap> I would imagine bitmap to be much faster, as no hasing needed
[14:49] <patdk-lap> more geared for matching against local network
[15:18] <drab> patdk-lap: that's a good point, I hadn't thoguht of it, thank you
[15:51] <SupaYoshi> Hi
[15:52] <SupaYoshi> How do I upgrade my 12.04-LTS server to 16.04 LOTS
[15:52] <SupaYoshi> *LTS
[15:58] <compdoc> I just didnt that for a system. worked great, and I was doing it over ssh
[16:00] <compdoc> oh, wait. I think it was 14.04. nm
[17:34] <lucidguy> is it possible to rsnapshop backup to a windows/ntfs share?  I would think not since it relies on hard links.. no?
[18:50] <xpistos> Hey all. I am having a bit of trouble with vim. I have several lines in the script that are formated mm-dd-yyyy like 01-01-2017, 01-08-2017 etc. how can I delete them all in one shot? I have tried g/^0?-*/d but that does not seem to work.
[18:51] <xpistos> And now that I think about it I will have like 10, 11 and 12 so even using ^0?-* wouldn't work anyway.
[20:11] <jancoow> Hi there. How to change this service into the new systemctl ? http://pastebin.com/8cnuJzZG
[21:22] <Doow> Hi! I'm trying to set up some backup scripts and I noticed that the backup user don't have write access to its home directory (/var/backups) What's the reasoning behind this? What's the preferred workaround (my specific problem right now is that gpg can't create the .gnupg directory)
[21:23] <patdk-lap> heh?
[21:23] <patdk-lap> backup user?
[21:23] <patdk-lap> there is no backup user unless you created one
[21:24] <patdk-lap> and /var/backups isn't exactly something you want to backup, as it's outdated infomation, why it's called backups
[21:25] <Doow> patdk-lap: I don't want to backup the contents of /var/backups, I want somewhere for my backup user to write metadata :)
[21:26] <Doow> patdk-lap: are you 100% sure there's not a backup user by default? cause I have no memory of creating one and a 'random' person in #ubuntu had the exact same settings as me in his /etc/passwd
[21:26] <Doow> backup:x:34:34:backup:/var/backups:/usr/sbin/nologin
[21:27] <patdk-lap> hmm, I do have a backups user on this system
[21:30] <Doow> I could of course create a "backups2" user with its own home directory that I control, but it feels a bit silly. I don't want to change the current user without knowing why it's setup the way it is.
[21:30] <Doow> In case I break something
[21:35] <jancoow> backupgs "sigh"
[21:42] <Doow> It looks to be a heritage from debian, they appearantly have an open bug from 2001 about it...