| Guest44 | Hi, does the devuan installer have options to install no desktop environment at all? I assume it would, considering the debian net installer does, but had to check in here first before considering switching to devuan. Thanks in advance if anyone answers my question (: | 10:58 |
|---|---|---|
| gnarface | Guest44: yes, as long as you use the right one. there's several installers, just don't use the live iso, i think the others should give you options | 11:02 |
| Guest44 | gnarface: thanks! | 11:03 |
| gnarface | no problem | 11:04 |
| gnarface | just to be clear, the devuan netinstaller is actually based on the debian one | 11:05 |
| gnarface | it should look very familiar to you | 11:05 |
| Guest44 | Awesome, good to know. | 11:05 |
| sfox | gnarface, i thought you would be interested to know, | 19:53 |
| sfox | there are two ways to do remote Xserver. One is to configure X to listen on inet6, generate a magic copy and copy it over. then you can run programs remotely from another system by editing the DISPLAY envvar | 19:54 |
| sfox | the other seemingly more fleshed out way is to configure XDMCP on the system you want to be used remotely (opposite of the machine running the Xserver) and then configure your thinclients to start X with X :1 -query bigcomputer.lan | 19:55 |
| sfox | It is actually suprisingly fast, to the point it was running my gecko based web brower on a thinclient and I couldn't tell the graphics were being done over the network | 19:56 |
| sfox | with a gigabit link | 19:56 |
| sfox | I'm trying to setup an NFS server, but portmap is missing it's config file and init script | 22:05 |
| sfox | https://wiki.debian.org/NFSServerSetup | 22:05 |
| sfox | is this guide still up to date? | 22:06 |
| sfox | /etc/default/portmap and /etc/init.d/portmap are missing | 22:06 |
| plasma41 | sfox: Given that portmap is now a virtual package provided by rpcbind, I'd say that guide is at least somewhat out of date. | 22:14 |
| sfox | is there better documentation that you know of for setting up nfs? | 22:17 |
| plasma41 | sfox: Sorry, no. I've not personally used NFS. Someone else might know more. | 22:18 |
| sfox | ok | 22:19 |
| rrq | sfox: after installing nfs-kernel-server there's good info in "man nfs" | 22:20 |
| plasma41 | sfox: Please do update that wiki page if you are able to figure out how to make it work. | 22:21 |
| rrq | ... and "man exportfs" has words on additional admin hands-on | 22:28 |
| rwp | I just last week powered down my Stretch NFS server. I'll power it back up and look at how it /was/ configured for you. | 22:29 |
| rrq | though mostly it's just set up in the /etc/exports file plus perhaps some firewalling to shield it a bit | 22:30 |
| sfox | nevermind i think i got it | 22:32 |
| rwp | The init script /etc/init.d/nfs-kernel-server is what starts everything at boot time. It has been included in the nfs-kernel-server package. | 22:32 |
| sfox | the problem was when i first started nfs-kernel-server it didn't start helper daemons because there were no exports yet | 22:32 |
| sfox | i set an export and service nfs-kernel-server reload | 22:32 |
| sfox | this had a nonzero exist status but did not start the nfs helper daemons like mountd | 22:32 |
| sfox | it's just a weird edge case | 22:32 |
| rwp | Yes. Something needs to be configured in /etc/exports first. | 22:33 |
| sfox | i did service nfs-kernel-server restart and was able to mount on my laptop | 22:33 |
| rwp | I was worried that they had removed the init script from the package. There is a lot of that going around. And then it must get added to the orphaned init script package. | 22:33 |
| rwp | It's good to see that is has not been removed yet. (I just checked for myself. Though I know you already said that.) | 22:35 |
| sfox | yeah that's what i was worried about too | 22:36 |
| sfox | but yay now i can listen to my music on my laptop and share it with my roomates | 22:36 |
| sfox | should work for movies too | 22:36 |
| sfox | I have it allowing my ipv6 GUA address so it shouldn't be routable to the internet | 22:39 |
| sfox | only within my site | 22:39 |
| rwp | NFS works well. But do remember to umount the mount point before removing your laptop from the LAN and taking it elsewhere or it will resume without connection to a file system mount point and that will appear as an error to the kernel. | 22:39 |
| sfox | thanks for the help | 22:39 |
| sfox | oh | 22:39 |
| rwp | If that happens use the -l,--lazy umount option. That should allow it to be unmounted. Pretty sure. I think. | 22:40 |
| sfox | could Mobile-IPv6 help with that? Is there a guide for setting that up? | 22:40 |
| rwp | This has no relation to IPv4/IPv6 differences. It's purely a matter of the kernel thinking that one of it's disk drives has departed the universe. | 22:40 |
| rwp | If you are never writing to the mount point then mounting it ro read-only is somewhat better because then there can never be dirty pages queued to write to the mount point. Usually always unmounts cleanly then. | 22:42 |
| sfox | IPv6 has a feature called Mobile-ipv6 which allows you to retain your IP address and connectivity when roaming to a foreign network | 22:42 |
| sfox | this should solve that problem right? | 22:42 |
| rwp | I would never use NFS over the Internet though. It's really meant as a LAN thing. Though I am sure someone will mention that NFSv4 using TCP is better about this. | 22:43 |
| sfox | is it using UDP by default? | 22:43 |
| rwp | That depends upon the version of NFS. So, it depends. And it depends upon the setup. | 22:43 |
| rwp | Also all network files systems suffer from high latency. And over the Internet all file systems suffer, suffer, suffer from the latency that is not there on a LAN. | 22:44 |
| rwp | Over the Internet I suggest using sshfs instead. And since that is a userland fusermount program (nfs is in the kernel) it means you can always kill and restart it and no problems. | 22:44 |
| sfox | okay. | 22:45 |
| sfox | it just sucks I have to do something differently depending on where I am | 22:45 |
| rwp | If you were using sshfs over mobile-ipv6 and maintained the same addresses so the tcp connection persisted then I could see that working while roaming. | 22:45 |
| rwp | Also if you used a VPN to use a persistent IP address connection then I could see that working while roaming too. | 22:45 |
| rwp | Try sshfs and see how it works for you. Then if you want to do it while roaming from home to remote and back I myself would use a VPN but that's a whole another layer of stuff to understand. | 22:47 |
| rwp | NFS is the highest performing network file system on the LAN though and that is why everyone needing high performance computing is using it rather than other things. But honestly sshfs is pretty good for just one person doing things like playing movies and such. | 22:48 |
| rwp | NFS client is in the kernel and the in-kernel client is well known to sometimes get wedged and stuck in such a way that the only way to resolve the problem is to reboot. It's due to when things change on the NFS server side. That one very small detail is almost the entire reason people love to hate upon NFS. Because things can get into a state that umount'ing does not work to clear the error and one must reboot. | 22:50 |
| rrq | is the 'soft' mount option any good against that? | 22:53 |
| rwp | "soft" has the problem of silent data corruption. | 22:56 |
| rwp | This has some good description of the tradeoffs for soft: https://kb.netapp.com/onprem/ontap/da/NAS/What_are_the_differences_between_hard_mount_and_soft_mount | 22:58 |
| rwp | I would say never use the soft mount option. But never for me might be a use case for someone. I won't say never. But I can't imagine a situation where I would want to use "soft". Though I am sure there is a use case somewhere. I just can't imagine I would ever want to do it. | 22:59 |
| rwp | And when I say silent data corruption it is not that I have read it from the bullet point of the marketing. I have seen people try to use soft and then we actually did have silent data corruption. People are always experimenting to try to improve performance or to fix some problem. And I say gaining the experience of actually trying things is a good thing. | 23:01 |
| rwp | But the result was silent data corruption. It's a real thing. And in my experience almost 100% likely to occur when "soft" is in use and things get busy. So I would never use it. | 23:01 |
| rwp | I should mention explicitly that I would use "intr" to make things interruptible. That's definitely a good thing. | 23:07 |
| sfox | libgtk2-perl was removed from Debian. Is there another place I can obtain it? | 23:15 |
| fsmithred | sfox, if libgtk3-perl won't work then I guess you can get it from oldstable. | 23:17 |
| fsmithred | or old-oldstable | 23:18 |
| sfox | gmusicbrowser : Depends: libgtk2-perl but it is not installable | 23:19 |
| sfox | I don't want gtk3, it's broken and horrible. | 23:19 |
| sfox | fsmithred, how would i do that? | 23:19 |
| fsmithred | go to packages.debian.org and download all the debs you need. If you still have a beowulf installation you could 'apt download' them | 23:20 |
| sfox | i found this | 23:21 |
| sfox | https://pkginfo.devuan.org/cgi-bin/package-query.html?c=package&q=libgtk2-perl=2:1.24992-1+b2 | 23:21 |
| sfox | can I just install the version from beowulf on daedalus? | 23:21 |
| fsmithred | you can try. | 23:22 |
| fsmithred | It may or may not work. | 23:22 |
| sfox | yeah it breaks apt | 23:32 |
| sfox | i'm going to try building gmusicbrowser from source and installing Gtk2::TrayIcon from cpan | 23:33 |
| rwp | libgtk2-perl remove ticket https://bugs.debian.org/986109 | 23:36 |
| sfox | there's no software to replace it | 23:48 |
| sfox | why should the GNOME team be allowed to sabotage other people's installations? | 23:52 |
| sfox | " GNOME team's goal of reducing gtk2 use in Debian" | 23:52 |
| sfox | cpan force install Gtk2::TrayIcon works | 23:56 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!