[00:00] alright, so now I just change sdb to xfs using the command above, move the tileserver to it and I should be golden, right? [00:00] it works the same, just on a different level/size [00:01] Glorfindel: have you just created a VG or have you also created an LV and put a filesystem on it? [00:01] ehhhhh what [00:02] this reminds me about the CERN data storage where the fast layer is (was?) spinning rust and the slow layer was tape robots... [00:02] Glorfindel: pastebin the commands you used [00:02] !pastebin [00:02] For posting multi-line texts into the channel, please use https://paste.ubuntu.com | To post !screenshots use https://imgur.com/ !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [00:03] I created the .img like you said, but haven't changed it to xfs yet or mounted it. then I mentioned that I added block storage and I'm assuming I can change the fs on that to xfs and run the tileserver from it, leaving the rest of the system intact? [00:03] Glorfindel: you don't need that img file now - you have sdb [00:03] yeah, if you mount that filesystem on the block storage in the right place [00:04] you want the tiles to end up on that :) [00:04] or maybe point the tile server to the right place, dunno what's easier [00:05] right, ok. so now I still need to set the filesystem to xfs, and move the data for tileserver over to it? [00:05] how do you get the data for the tile server? [00:06] rsync? [00:06] I run java -jar ./tileserver/tileserver.jar [00:06] tileserver.jar and the config files I got from github [00:07] I was planning on cp-ing them over [00:07] Glorfindel: anyway - just start off with 'vgcreate myfineandgloriousvolumegroup /dev/sdb' and then 'lvcreate -n mynewlogicalvolume -l 100%FREE myfineandgloriousvolumegroup' and then mkfs -t xfs /dev/myfineandgloriousvolumegroup/mynewlogicalvolume [00:07] then mount it [00:08] Glorfindel: and use rsync -avPHAX instead of cp [00:08] is it ok if I change the name of my volume group? :P [00:08] obviously [00:08] :) [00:08] and the name of the LV, of course [00:10] mount: can't find /dev/sdb in /etc/fstab [00:11] running "mount /dev/sdb" [00:12] https://paste.ubuntu.com/p/Q2vV3wpQQ6/ [00:13] ignore everything after the "ls /dev/" line [00:13] well, obviously you'll need to run those lvm and mount commands as root [00:14] and you'll need to mkfs -t ext4 /dev/minetestvg/tileserver and that's the one to mount - not sdb [00:15] also - to have it mounted automatically, add it to /etc/fstab [00:16] didn't I want to create xfs though? [00:16] or is ext4 part of xfs [00:16] two separate things [00:16] ok [00:16] there's no mkfs.xfs or mkfs -t xfs in that pastebin, and you're trying to mount the pv, not the lv [00:17] RoyK: wait why ext4 here? [00:17] sarnold: sorry - my fault [00:17] Glorfindel: no - don't put ext4 on that - put xfs on it - typo [00:17] tiredtypo [00:18] too late. rerun it with xfs? [00:18] yep [00:18] it'll tell you there's an fs there [00:18] mkfs: failed to execute mkfs.xfs: No such file or directory [00:18] apt install xfsprogs [00:20] https://paste.ubuntu.com/p/Z2PGJTqP35/ [00:20] am I supposed to put -f all alone? [00:20] rather than with the -t [00:20] -ft xfs would probably work [00:21] no -f [00:21] if not, then indeed -f -t xfs [00:21] sarnold++ [00:21] -t takes an argument [00:21] -f does not so it has to come first [00:22] invalid option 'f' [00:22] idgi [00:22] pastebin that command [00:22] on it [00:23] https://paste.ubuntu.com/p/g8CqG7VTm5/ [00:23] erm - move -f to the back or just run mkfs.xfs directly [00:23] mkfs -t xfs is an alias for mkfs.xfs [00:24] ow. I sure didn't expect *that* mess.. [00:25] mkfs.xfs -f /dev/restof/path worked [00:25] sarnold: well, you get what you pay for and it's free ;) [00:25] RoyK :D [00:27] ok, so attempting to mount it, still complains of not being able to find it in /etc/fstab [00:27] mount /dev/bla/bla /somewhere [00:27] ooh I get it [00:28] then - when it works - add it to /etc/fstab [00:29] ok, I'll see if I can figure out this stuff myself, seems simple enough [00:29] after adding it there, umount /somewhere and try 'mount -a' [00:29] it should work [00:33] worked [00:33] :) [00:33] it seems I need sudo to modify it though, how do I fix that? [00:33] fix what? [00:34] you need sudo to do system things [00:34] needing sudo to add a directory [00:34] so I'm assuming I'll need to run the .jar as sudo also [00:34] depends on where [00:34] !chmod [00:34] An explanation of what file permissions are and how they can be manipulated can be found at https://help.ubuntu.com/community/FilePermissions [00:34] !chown [00:35] oh, right. I'll investigate chown then. thanks :D [00:49] RoyK: everything seems to be working perfectly, many thanks for walking me through all that! [00:49] nice :) [00:50] once I open the port I'll share a link to what I was running [01:04] Glorfindel: good :) [01:25] eh, I need to mess around with this a bit more before it runs I guess. oh well. Thank you all and have a good night :) [01:28] but here's it running on someone elses map https://pandorabox.io/map/#-1693/-58/8 [01:29] that's cool [09:33] 18.04 (server) has an /etc/sysctl.d/xxxx that contains "net.ipv4.ip_forward=1" but it isn't being set ("systemctl status systemd-sysctl") shows the service ran. How to debug? "/lib/systemd/systemd-sysctl --prefix=net.ipv4.ip_forward" wants to set it correctly [11:27] Hello, i faced with problem i was run ZNC but when i try connect to ZNC i got error (connection refused), what can be problem? P.S. znc online. [11:39] coreycb: jamespage: can you please remind me of the URL for the version check matrix for UCA packages? somehow this has dropped from my list of bookmarks [11:41] good morning [11:41] frickler: https://wiki.ubuntu.com/OpenStack/VersionTracking [11:43] coreycb: perfect, thank you [11:52] Kazaaak: sounds like znc is no longer running or no longer reachable on the ip:port it binds to from where you are trying to connect from. [12:11] znc true that running because i see nick is online, by /whois and i try connect via mIRC by command /server ... and i not sure or at server are firewall maybe you can tel what command at terminal i can open ZNC port? [12:17] Kazaaak: just running znc should make it bind to a port which you can then connect to. if this doesn't work, check your firewall configuration, try to connect with nc -v from the same server and from remote to see whether the TCP connection can be established. === emerson is now known as GDPR [13:05] Good time, [13:05] I'm here too help Kazaaak get help.. [13:08] so we problem is that we can't connect to znc via mIRC, i think we need open port at ubuntu terminal window, so who can tell what command line can open port if port is 1234 [13:08] I have an old sysv initscript (squid on xenial) that has a restart) action, and does some checks there [13:08] ? [13:08] xenial uses systemd, though, so that is "converted" into a service [13:08] but it looks like systemd has no specific restart action and always interprets that as stop + start [13:09] which means the code in the restart() function of the sysv script is not executed. It looks like systemd just calls the sysv script with stop and then tsart [13:09] is that interpretation correct? [13:09] (my interpretation) [13:14] cpaelzer: do you know? ^ [13:17] looks like it's a known issue: https://github.com/systemd/systemd/issues/2175 [13:21] ahasenack: yes afaik it is as you describe [13:28] ok [13:35] Has ubuntu changed how initd scripts are generated from the debian dir between xenial and bionic? I'm looking at debian/sssd-common.sssd.init.in and I can't find the file anywhere in xenial, even though it was introduced in 2010 [13:36] check d/changelog for mentions, and also d/rules to see how it is and was installed before [13:36] bionic has the file exactly in /etc/init.d/sssd, and xenial has something vaguely similar in /etc/init/sssd.conf, but it's not the same file [13:36] it might have been converted to systemd [13:36] or upstart [13:36] /etc/init/ is an upstart config directory [13:37] ok [13:37] you know, you can never have enough init systems [13:37] so if I wanted to backport pidfile handling stuff, I'd need to redo it according to upstart conventions? [13:37] looks like it [13:51] tomreyn, thank you that you try help, but i dont know how to check firewall with nc -v, i need try write at terminal at same server where running znc this command: nc -v :? Then from another server terminal same command? [16:29] is there any official documentation and best practice for using network namespaces on ubuntu 18.04? does netplan support network namespaces? [16:49] Is it possible to do uefi, preseed and software raid 1 with 14.04? [16:53] kierank: never tried with preseed - but then - why 14.04? it's soon out of support [16:54] just to match some existing configurations [16:54] kierank: preseed is certainly supported in 14.04. [16:54] yes we do preseed and software 1 raid one with legacy bios [16:54] but not uefi [16:54] kierank: but as RoyK said, all of those existing configs are going eol in ~1 year [16:54] kierank: oh i see what you are asking now, sorry [16:55] nacc: or ~7 months :þ [16:56] RoyK: rounding up ;) [16:56] kierank: if it's not too complex stuff you're working with, an upgrade should be trivial [16:56] yes we plan to update soon but we need to match some existing configurations in a closed environment [16:59] trusty is likely to get an ESM service too; if you need to keep running it beyond EOL, have a conversation with an ubuntu advantage support person about your needs [17:00] sure [18:09] sarnold: announced a little over a month ago: https://blog.ubuntu.com/2018/09/19/extended-security-maintenance-ubuntu-14-04-trusty-tahr [18:09] nacc: oh :) nice [18:10] nacc: thanks [18:15] I can't find how long is ESM in that page [18:15] seem like the #1 selling point for it :) [18:23] sarnold: np; tbh, I wasn't sure it was announced or not, but I thought it was, so I was checking for my own edification [18:24] that was a sprint week for me, I'm not at all surprised it didn't make an impression on me :) [18:24] heh [18:39] How to check or at server are firewal? [18:40] Kazaaak: Not sure I parse that, you want to check if there is a firewall on a server? [18:41] "iptables-save; ip6tables-save" should tell you what's loaded in terms of ruleset [18:52] nacc, yes, i want to know or firewall blocks one port, because i was run ZNC and i can't connect via mIRC. Can you help me? [18:54] Kazaaak: as sdeziel said, check with iptables and ip6tables, either those save commands above, or iptables -vnL and ip6tables -vnL [18:59] has anyone used keepalived+conntrackd+radvd (or different combo) to setup HA firewalls? I'm looking for some advises to address a problem during failovers [19:02] sdeziel: not with firewalls, but with a postgresql install - should be about the same, except some TCP sessions will be lost [19:03] didn't use conntrackd, though [19:03] RoyK: interesting, since mine are firewall, they only see TCP sessions cross them and conntrackd makes the state sync between the 2 nodes [19:03] didn't know that could be clustered [19:03] RoyK: yeah, works very well [19:04] what issues are you getting? [19:04] I see "bogus" RA being sent to clients [19:04] I have keepalived setup to use fe80::1 as VIP and that's what clients initially pick up as their default gw [19:04] all IPv6 or mixed? [19:04] dual-stack but v4 works flawlessly [19:05] when I restart keepalived it moves the master role and along with it radvd [19:05] ok - perhaps file a bug, then - maybe someone hasn't really worked too hard on the ipv6 support [19:05] Good, RoyK, i wanna check only IPv4 so i can enter command to terminal window: iptables -vnL ,and what i need sea then? [19:06] but doing so result in the clients picking up the link local IP of the new firewall as another gw so they have 2 gw and they load balance between both which I don't want [19:06] RoyK: I'm sure I'm the one at fault here ;) [19:11] whats a good package for sharing one monitor with multiple servers? Is synergy still the only solution ? [19:12] hardware kvm boxes are popular [19:12] dunno if they are *good* but they're popular [19:13] sarnold: true, but i used to use Synergy, network based app, no hardware need, and it did what I wanted to do.. trying to reduce number of devices around me :) [19:14] normally when you have to use a monitor something has gone wrong.. it's nice to have the simplest possible tools at that moment ;) [19:15] indeed [19:18] tomvolek: My favourite answer is a network-accessible hardware power/kvm switch. Retrofits to existing boxes and makes them act like enterprise server hardware. [19:19] @mason any brand/model in mind ? [19:20] tomvolek: I'm trying to remember what I used to use, but failing that I'll find one of the same class. [19:20] tx, will google it [19:25] tomvolek: It's possible the vendor I used is gone now - it was much of a decade ago. Something vaguely like this: https://www.startech.com/Server-Management/KVM-Switches/4-Port-USB-VGA-IP-KVM-Switch-with-Virtual-Media~SV441DUSBI [19:25] The ones I deployed last did both KVM and power. [19:25] I think i need just try open port and be see or it's helps or not, can anyone tell what command i can open port: 1235 on Ubuntu 18.10 ? [19:26] Kazaaak: the commands to use to open it depend upon the commands you used to close it in the first place.. [19:26] Kazaaak: could you pastebin the iptables-save output as well as "ss -nlt"? [19:29] tomvolek: I'm going to give up as I'm finding neither email where I talked about it nor the actual product, and I'm not remembering the brand, but it was something like https://www.startech.com/support/SV1108IPPOW , only with more ports. [19:30] @mason thanks my friend, dont worry, i will find something, much appreciate it [19:30] tomvolek: Sure. Enjoy. [19:44] sarnold, i don't close the port, i just install ZNC and i can't connect, i think at my server is on firewall... I don't know:/ [19:44] sdeziel, afcourse i can give you output at pastebin, jus tell me exactly what i need enter input command? [19:45] jamespage: ping re the LP team ubuntu-server-iso-dev [19:46] Kazaaak: ( iptables-save; echo; ss -nlt ) | pastebinit [19:51] hggdh: wow does that still exist? [19:51] * jamespage waves at hggdh [19:52] jamespage: hi, long time :-) [19:52] hggdh: indeed [19:52] jamespage: yes, it is still there, and there are 4 or five pending joins. Looking at it, and the ML, it seems mostly dead [19:53] jamespage: should we go ahead and simply bury it? [19:53] hggdh: +1 I think it got superceeded a while back [19:54] jamespage: what is the replacement (so that I can let these applicants know) [19:54] that is a very good question [19:54] heh [19:55] I will go ahead and announce the funeral, and refuse the applicants; then I will find out how we can kill it, effectively [19:56] Ok lets try, sdeziel, little bit wait. [19:59] hggdh: I think this replaced it but it may have since had a similar funeral - https://utah.readthedocs.io/en/latest/index.html [20:00] yeah [20:04] jamespage: renewall is set to "ask" instead of auto-renewal. I set a note in the home page stating this team is no longer active, and declined the pending candidates (most requested around 2013) [20:06] sdeziel, https://www.irccloud.com/pastebin/afeqOwDm/ [20:07] Kazaaak: is that a Ubuntu server machine? (also, please avoid pastebin services requiring javascript) [20:17] sdeziel, yes Ubuntu 18.10 VNC [20:17] an ubuntu machine without iptables or iproute2? o_O very curious [20:18] A box without hinges, key, or lid, yet golden data inside is hid. [20:18] my precious! [20:18] Kazaaak: you don't seem to have a firewall configured [20:18] and yeah, no iproute2 is weird [20:19] but I haven't encounter that many cosmic machines [20:34] sdeziel, yes stranger. My admins rent me server from appboxes.com.. I can't connect to ZNC port what i run. [20:35] is there a subiquity bug for "Use FUll Disk and set up LVM" where it doesn't take more than 4GB of space for the LVM? [20:36] teward: yes [20:36] what bug number is it? [20:36] so I can subscribe and track [20:36] teward: https://bugs.launchpad.net/subiquity/+bug/1785321 [20:36] Launchpad bug 1785321 in subiquity "LVM Entire Disk option does not use entire disk" [Undecided,New] [20:40] ahasenack: sounds to me like we need to make some changes to subiquity then to 'ask' how much of the LVM to utilize. because 4GB is... real small. [20:40] was this 'change' documented in the release notes? [20:41] or did I miss that too? [20:41] I heard arguments that it was correct behavior, in the sense that the VG is using the whole disk, and the LV is easily resizeable [20:41] but I also think it's a bug [20:41] i'm on the fence. [20:41] teward: it's a bug, not a "change", so no release notes entry [20:42] if you use ext4 for /, you can even resize it without booting into rescue mode [20:42] or so I heard [20:42] ahasenack: yeah, i'm thinking if we do anything that isn't 'the whole VG' for the LV, then we should be doing 50% of the LV. LVM was kind of the 'thing' this became, and since it's a VM it's possible I'll expand this over multiple virtual harddrives within VMware spanning multiple datastores if this gets large. [20:43] but that's just me :P [20:43] *goes back to fixing things now that he's expanded the LV to fill up the VG* [20:43] reducing the LV after installation is more annoying [20:43] but 4Gb by default is unexpected, that will easily be filled [20:44] well, then at least something larger than 4GB [20:44] 10GB even is better than 4 [20:44] yep [20:44] chime in over there :) [20:44] can't from my phone but will later when I get home probably. [20:44] hehe [20:44] understood === ShaRose_ is now known as ShaRose