11:32:52 Smithx10: what is pixiecore? 11:38:08 Found it 11:39:08 https://github.com/danderson/netboot/tree/main/pixiecore 12:15:56 papertigers: I'm able to reproduce that. If I set "vfs modules =" in the share, i.e. disabling the acl modules, then it appears to work. 12:16:20 dtracing fsp_get_io_fd shows its coming through the solarisacl module, so I'm building a version of that with the same changes as that bug report 12:17:51 given https://github.com/NetBSD/pkgsrc/blob/trunk/net/samba4/patches/patch-source3_modules_vfs__solarisacl.c I expect nobody uses that and it's just been missed 14:00:57 nahamu: you working over at equinox? Using tinkerbell? 14:18:41 randomly I'm experiencing "initial termio on /dev/wscons failed" ...as I can see the problem was reported some times in the past but never solved ...do you have any hints? Here is also the screenshot https://ibb.co/c3Bq1R9 14:52:11 it is possible that streams stack setup is not completed on wscons 17:15:59 jperkin: sorry I didn't file anything, I was hoping to do a little digging to make sure I didn't do something dumb on my end 17:16:11 It looks like there's a real issue there then? 17:19:40 yeh, I kinda got sidetracked by being unable to dtrace userland function arguments that are being loaded from modules, binaries and shared libs show them but not the samba .so's, but they have CTF data... 17:20:04 not sure what's going on there, but otherwise am looking through to see what's possibly been missed in the solaris acl module 17:59:18 jperkin: adding that line to my share seems to have made the error in the log file go away but I cannot see the files still. I will play around with it tonight if I have time 18:04:31 jperkin: nvm I still see the log message.. where in the smb.conf did you stick "vfs modules =" 18:19:21 yeh ignore that, I thought it was working but it wasn't 18:28:03 do you see the same behavior though? Where you can see the directories and subdirectories but not individual files? 18:28:38 Do you keep older pkgsrc trunk builds around anywhere that I could downgrade to, or are they cached locally? 18:30:28 yeh I see the same, at least from a macOS client. I tried the 2023Q4 package but that fails too. I don't keep old trunk builds, but may bisect the pkgsrc updates 18:48:41 and old tar.gz files for packages don't hang around on the filesystem anywhere? 18:56:55 I don't think our mirror does cleanups so we might still have some older builds. Oldest for trunk is samba-4.15.6nb3.tgz from 07-Jul-2022 if that helps 19:02:26 wiedi: could you push that somewhere or can I just hit pkgsrc.smartos.org 19:02:42 https://packages.pkgsrc.pub/smartos/SmartOS/trunk/x86_64/All/ 19:09:22 jperkin: if it helps at all this version works 19:09:23 https://packages.pkgsrc.pub/smartos/SmartOS/trunk/x86_64/All/samba-4.18.8nb3.tgz 19:09:30 wiedi: thanks! 19:11:21 Another weird bit of info I can provide is that my Sonos system was still able to play music in a playlist from the samba share when it already knew about the full file path. It seems that listing was the major thing that was broken 19:28:05 wiedi: oh that's helpful, thanks ;) 19:29:20 papertigers: ok, good to know, I'll dig in more tomorrow 21:45:33 Smithx10: I am indeed at Equinix, sadly Tinkerbell never made it to production, but we do still use and contribute to the Tinkerbell ipxedust library. 21:54:31 @nahamu yea, I had to write something in house using go to do our initrd / booting linux with an opinion around zfs 21:54:46 Was annoying lol