=== mundus2018 is now known as mundus [09:51] Hi Everyone, setting up a home 3 disk array (running / on the array), for partitioning would a seperate /boot partition (raid 1), the / partition (raid 5) and the swap partition (raid 0) be an acceptible scheme? (and not cause problems resilvering a disk after faliure) thanks :) [09:53] stripe: why the different RAID levels per FS, if there's the same 3 disks underlying it all? [09:55] TJ the /boot is raid 1 to allow booting from any drive after faliure, the swap (0) is to minimise the amount of writes. or am I over thinking it (as usual lol) [09:57] stripe: I wouldn't like to be using RAID-5; it is very risky especially with 3 disks [09:58] thanks TJ what would you use? [09:58] once you lose a disk you've no resiliance until you've replaced it with a hot spare, and most admins will tell you with RAID-5 *don't touch the array, don't shutdown" until you've got the replacement drive in place (oh, and do a full backup BEFORE adding the replacement disk) ! [10:03] TJ would you add another spare disk to the array or run a 4 disk (raid 5) array? I will have all my data snapshot'd and backed up. [10:09] I wouldn't use RAID-5 at all; RAID-6 or RAID-10 (mirror + stripe) [10:10] stripe: are these real spining disks or actually SSD ? [10:12] stripe: short informative read here: https://storageswiss.com/2015/01/13/which-raid-level-use-for-ssd-tier/ [10:21] thank you TJ will look at 6 and 10, all spinning drives, and thanks for the link. all the best :) [10:23] stripe: I'd always prefer RAID-10 even if it does give less capacity than 5/6 because it so much easier to rationalise about and work with when a disk (or two) fail [10:33] TJ the loss of capacity is not an issue, (will mainly be running LXD{lxc} containers) so will do some further reading, thanks again for the help. have a great day :) [10:34] GB are cheap. What RAID10 loses in $$ for storage space, it gains with simplicity and robustness. [10:35] thanks blackflow, my two favourite words simplicity and robustness it looks like I will be using raid 10 :) [10:35] also btw I wouldn't recommend RAID0 for swap. In case of disk failure it's crashy-crashy time, and the whole point of RAID is to keep the server running while you resilver the faulty disk [10:36] (and if at all possible, use ZFS) [10:38] blackflow, have used zfs on freebsd, never even thought about using it on linux (facepalm) excelent reminder :) [10:39] d'oh! [10:39] :) [10:41] stripe: the only "problem" (for now) is that you'll have to install Ubuntu from debootstrap, until the installer grows the ZFS capabilities, which is in progress. But hey, I prefer deboostrapping anyway, I end up with most minimalist possible installation, no cloud-init, snapd and similar bloatware. [10:45] debootstrap is not a problem, thats how I originally built my containers/chroot's (back in the day) [10:46] ZFS it is, then :) [10:50] yep :) [11:10] blackflow: the only issue with zfs is the lack of flexibility [11:11] blackflow: you can't just toss in a new drive in a VDEV or remove one - it's rather static on the VDEV level - there are talks of this changing, but I haven't seen the code yet [11:14] precisely why a 2-disk mirror vdev is best. you just keep expanding by adding 2 disks at a time (or replacing them with bigger ones) [11:15] 2-disk mirror vdev, and then many vdevs are "stripes", so it's effectively a RAID10 as ZFS doesn't explicitly have a "RAID10" option [11:28] blackflow: for striped mirrors, yes, but I was thinking of raidz things [11:32] yeah. [15:47] bryce: bug 1773324 came up in triage. Are you OK to re-review/sponsor please? [15:47] bug 1773324 in rabbitmq-server (Ubuntu Eoan) "rabbitmqadmin shows %%VSN%% as version" [Medium,Triaged] https://launchpad.net/bugs/1773324 [15:49] bryce: and nice job helping the volunteer to get the patch into shape :) [16:28] rbasak, sure will take a look at it now [23:11] Hello! [23:11] Could someone tell me if Ubuntu Server comes with an active firewall nowadays? [23:13] Please? [23:14] supercool: ufw is available to install if you want a friendly front end to iptables [23:15] sarnold: actually I am just trying to figure why am an unable to access a server from a client server [23:17] I think iptables are not initialized here [23:18] So, no firewall? [23:18] how about nftables or ebtables? [23:18] I don't think there is a firewall config blocking SSH or HTTP etc by default. probably no significant firewall by default [23:19] I am trying to use 8000 and 8080 port and none of them worked [23:19] I mean couldn't get anything from host side [23:19] are you on the same lan, the same subnet? [23:20] how do you know some application is listening on those ports? [23:20] ploxiln: I am running a Django server on those ports [23:21] I can acess from 127.0.0.1:8000 or 127.0.0.1:8080 using curl [23:21] is it listening on 0.0.0.0 or 127.0.0.1? [23:21] But the host doesn't receive anything [23:22] I bridged the guest, So I think it is listening on 0.0.0.0 from outside [23:22] so this is running in a VM on your laptop/desktop [23:22] and you are trying to access from the browser on your laptop/desktop? [23:23] My network IP would be 192.168.15.45:8000 for example [23:23] Yes, exactly [23:23] can you reach it on the non-loopback IP if you curl on the same box? [23:24] tds: yes [23:24] hmm, well, that might not do it, the VM might not have the same ip as your laptop. depends on VM setup. how do you ssh to the VM? [23:24] the laptop/desktop may have a firewall enabled by default [23:24] ssh user@192.168.15.45 [23:25] hmm. ok, I would expect that accessing http://192.168.14.45:8000 from the same context as you ssh would work [23:26] from the laptop/desktop, outside the VM [23:26] ploxiln: yes, but it is not [23:26] if run "ip addr" inside the vm, do you see the same or similar address in the output [23:27] inet 192.168.15.45/24 brd 192.168.15.255 scope global dynamic [23:29] what's the full output of `iptables-save`? [23:30] tds: nothing shows [23:30] or shows nothing* [23:30] ok, doesn't sound like firewall then [23:30] unless there's ebtables of nft rules as sarnold suggested, seems unlikely though [23:31] looks like django listens on 127.0.0.1 by default, you need to specify 0.0.0.0:8000 [23:31] https://code.djangoproject.com/ticket/27537 [23:31] Django bug 27537 in Core (Management commands) "Provide a simpler way to default runserver IP/port to 0.0.0.0:8000" [Normal,Closed] [23:33] ploxiln: you are right! o/ [23:33] sounds like it's already bound to the right ip if you can reach it from 192.168.15.45 on the same host though? [23:33] oh, weird [23:33] I had to runserver as 0:8000 [23:34] 0 is a shortcut for 0.0.0.0. Full docs for the development server can be found in the runserver reference. [23:34] good to hear :) [23:35] Thank you a lot! What is your account number for the deposit? [23:36] Thank you guys!!! [23:36] haha. uh, 42 [23:38] ploxiln: \o/