armadefuego | Just an FYI: | 07:19 |
---|---|---|
armadefuego | I suspect that we have a mismatch in how the data is being stored, and interpretated, in Zm. | 07:19 |
armadefuego | That may be the cause of some _missing_ videos. | 07:19 |
armadefuego | The full path includes that date, and the date is subject to UTC/local issues. | 07:19 |
armadefuego | The save is done on the kiosk, but the view is on the server ;{ | 07:19 |
armadefuego | for the most part the web service is php, so the php.ini should reflect the date base of the db saves. | 07:22 |
armadefuego | Even tho the zm version is the same the base package is different. ( 14.04 and 18.04) | 07:22 |
=== cpaelzer_ is now known as cpaelzer | ||
=== dbungert1 is now known as dbungert | ||
zph1nx | can i ask subquity questions here, or is that a separate channel? | 15:59 |
zph1nx | working with an ubuntu-server autoinstall, and it's been stalled here for the last half hour https://ibb.co/RQPmjY8 | 16:01 |
JanC | if it's a question about using it as part of a server install this place is fine | 16:02 |
zph1nx | lets see what disabling ipv6 in grub does | 17:27 |
zph1nx | funny how an install that takes 10 minutes on a shitty laptop takes x20 more time on a PowerEdge R760xd2 via idrac | 17:28 |
zph1nx | turns out it was ipv6 | 17:37 |
JanC | you were netbooting? | 18:01 |
JanC | or do you mean the kernel commandline? | 18:02 |
zph1nx | booting a user/meta-data from a remote webroot | 18:08 |
zph1nx | and yes, kernel commandline in grub.cfg | 18:11 |
trippeh | hrm after upgrading to 24.10 my fakeroot seems borked. fake-sysv is spending a loooot of time closing file descriptors that do not exist. like a billion of them. | 21:54 |
trippeh | oh well, downgrading fakeroot to noble versions did not help | 21:58 |
trippeh | maybe I broke something else. :) | 21:59 |
trippeh | oh seems like it is the vscode terminal setting open files ulimit 1073741816 | 22:01 |
patdk-lap | what did you do to open that many files in vscode? | 22:04 |
patdk-lap | open all / | 22:04 |
trippeh | ulimit != actually opened files. it probably just sets the soft limit = hard limit on startup | 22:04 |
patdk-lap | ulimit is actually number of open file handles | 22:05 |
patdk-lap | 1 handle per file | 22:05 |
patdk-lap | but 2 handles per tcp socket | 22:05 |
trippeh | it's the limit. hence uLIMIT :P | 22:05 |
patdk-lap | but then you also need a handle for like pipes, .... | 22:05 |
patdk-lap | yes, but if you hit the limit, you had atleast that many open file handles | 22:06 |
trippeh | anyway fake-sysv probably just tries to close any potentially inherited handles | 22:06 |
trippeh | they all return bad file descriptor | 22:06 |
trippeh | (any = anything not 0, 1 and 2 | 22:07 |
trippeh | ) | 22:07 |
patdk-lap | doesn't sound like ulimit | 22:09 |
patdk-lap | sounds more like apparmor to me | 22:09 |
trippeh | anyway setting 1024 fixed it | 22:09 |
patdk-lap | vscode or fakeroot? | 22:10 |
trippeh | ulimit -n 1024 in the vscode terminal before kicking off sbuild | 22:11 |
trippeh | doing a close(3..max open file ulimit) loop in utilities not wanting inherited fds is pretty common | 22:12 |
trippeh | or at least 3..1024 | 22:12 |
trippeh | trying to close a billion takes a while though | 22:13 |
=== nshirelaptop2 is now known as nshirelaptop |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!