IRC Logs for #crux-devel Saturday, 2018-07-21

*** pedja has quit IRC00:46
Romsterworking fine on my nvidia desktop now02:19
jaegerglad to hear it02:20
*** _________mavric6 has quit IRC02:48
*** _________mavric6 has joined #crux-devel02:49
jaegerAdded -sl versions to the crux-glvnd repo03:14
jaegerargh, and I messed them up03:15
jaegerwell, messed up a name03:16
jaegerok, fixed that03:20
*** xor29ah has quit IRC04:38
*** xor29ah has joined #crux-devel04:40
Romsterthe only reason i was using the -sl ports before was nvidia was lagging behind on xorg-server 1.20.x06:25
*** pedja has joined #crux-devel12:02
*** xx1973 has joined #crux-devel12:54
*** xx1973 has quit IRC12:56
*** xx1973 has joined #crux-devel12:56
*** xx1973 has quit IRC13:03
*** xx1973 has joined #crux-devel13:03
*** onodera has joined #crux-devel13:10
*** onodera has quit IRC13:11
*** onodera has joined #crux-devel13:12
*** xx1973 has quit IRC13:17
juejaeger: no problems with your mesa/libglvnd ports, using it with intel and nouveau13:19
jueno objections to push them with a notify commit13:20
jaegerThings like this make me think we should set up some kind of news feed on the website to which people can subscribe, and something they can easily fetch via command line13:44
jaegerBecause if we do switch to the glvnd stuff now, we'll have to answer questions about leftover symlinks for gl-select users, conflicts with mesa3d and libglvnd, etc.13:44
jaegerWe can send out a mail to the ML about it, of course, but I think some folks would still miss it13:45
juewell, I think the [notify] commit is a nice and easy to handly thing and the message is where it belongs: it's part of the commit. But, of courses, I've no objections if we offer an additional thing here, should be possible with an git hook like the one for the mail13:52
jaegerYeah. And I don't mind answering the questions, just thinking about something for the future that would be easy to see13:53
jaegerfrinnst: have you tried the glvnd stuff with your AMD GPU yet? I can't recall, sorry13:58
frinnstyeah works fine14:03
pedjaRomster, they are tested only with nvidia14:06
jaegerfrinnst: ok, cool. I don't have mine to test yet so appreciate it :)14:08
pedjafrinnst, no glibc breakage so far. nice job :)14:08
jaegeryeah, haven't experienced any here, either14:36
pedjaRomster, you might want to bump meson to 0.47.115:10
jaegerjue: oops, forgot to say thanks for the report in intel/nouveau :)15:20
*** groovy2shoes has quit IRC15:31
*** groovy2shoes has joined #crux-devel15:33
juejaeger: you're welcome :)16:02
jaegerhrmm, NFS mounts to my NAS seem to be broken all of a sudden, wonder what happened20:20
*** darfo has quit IRC20:23
jaegerany significant change to nfs-utils recently?20:24
*** darfo has joined #crux-devel20:25
jaegerAnyone currently running nfs on 3.4 who can maybe tell me what I'm missing?20:34
jaegerrpcinfo doesn't help, just gives: No remote programs registered.20:34
jaeger$ showmount -e20:46
jaegerclnt_create: RPC: Unable to receive20:46
jaegerno iptables rules in the way20:47
jaegerI can mount an NFS export from a CentOS VM on my crux system but can't do anything to share from the crux system anymore, hrmm21:01
jaegerno services are listening on active ports21:13
jaegerss | egrep -i '111|2049|rpc|nfs' <-- returns nothing21:13
jaegerer, ss -l -u | egrep -i '111|2049|rpc|nfs'21:14
jaegeron the centos VM I do see nfs and sunrpc outputs from that21:14
jaegerthis is weird, I can't find any useful info at all from the system. It just refuses to work21:19
jaegerjue: since you're the nfs maintainer, do you have any ideas?21:19
darfoglibc 2.27 has a patch that touches rpc. maybe the mounts timed out and now rpc is using the new glibc?21:43
darfodon't have an NFS setup to test with, sorry21:43
jaegerI considered that, have rebooted it21:51
jaegeroh, shit, you removed the multilib dirs patch22:03
jaegerfrinnst: is that stuff rolled into the big patch?22:07
jaegerI guess I can look myself22:07
jaegerAh, it is. Scared me for a bit :)22:08
pedjaif that isn't a case for a bit more detailed commit messages, I don't know what is :)22:15
frinnstI think we should create a glibc repo on to track everything22:45
jaegerdowngrading glibc doesn't seem to change anything, though I haven't rebuilt the rpc and nfs stuff22:45
jaegerNo objections here22:45
jaegerThis NFS thing is frustrating... it was working fine, now it isn't. I don't know which change or port update broke it23:00
*** onodera has quit IRC23:10
frinnstjaeger: I had some very odd nfs issue a few weeks back. was unable to mount23:26
frinnstrebooting the client didnt help. then i reloaded the iptables rules on the server and it all came back to me23:27
frinnsthas worked fine ever since23:27
jaegerweird. I have no iptables rules in place on the server at all :/23:27
jaegerThe NAS has a bonded interface but I've tested on 2 other crux 3.4 systems, one VM, one physical. Same results, can't serve NFS at all anymore23:29
jaegerclient is ok, server is hosed completely23:29
jaegerIf I run rpcbind manually it starts but has no output whatsoever, either stdout, stderr, OR syslog23:31
jaegerno listening ports23:31
jaegerooh, I think I finally found something23:39
jaegerrunning rpcbind manually and then stopping it with ctrl-c returned servname errors23:41
jaegeryeah, that's the problem23:45
jaegerjue: rpcbind needs this patch:
jaegeryeah, check this out:23:54
jaeger$ egrep 'sunrpc|rpcbind' /etc/services23:54
jaegersunrpc          111/tcp         rpcbind         # Remote Procedure Call23:54
jaegersunrpc          111/udp         rpcbind23:54
jaegeregrep 'sunrpc|rpcbind' /etc/services23:54
jaegersunrpc            111/tcp23:54
jaegersunrpc            111/udp23:54
jaegerthe new iana-etc services file doesn't contain the 'rpcbind' aliases23:54
jaegerso I suppose either patching rpcbind with the above patch OR modifying /etc/services to contain the aliases would work23:55
jaegerthe new services file has no aliases, though23:55
jaegerso I'd vote for patching rpcbind instead23:55

Generated by 2.14.0 by Marius Gedminas - find it at!