[04:44:37]
<nullitope> Hey folks, I'm running a Pleroma instance with Yunohost/BuyVM and upgraded my server a few days ago. I'm kind of lost on how exactly I do a server migration - my BuyVM page shows that the server is upgraded to the new amount of storage/memory, and when I open the console it starts up with Yunohost already installed, but when I run a diagnosis in the Yunohost web UI, it reflects the old amount of storage/memory. Do I need to reinstall Yunohost entirely, or what?
[04:45:07]
<Monitor Kernel Access> > <@nullitope:nitro.chat> Hey folks, I'm running a Pleroma instance with Yunohost/BuyVM and upgraded my server a few days ago. I'm kind of lost on how exactly I do a server migration - my BuyVM page shows that the server is upgraded to the new amount of storage/memory, and when I open the console it starts up with Yunohost already installed, but when I run a diagnosis in the web UI, it reflects the old amount of storage/memory. Do I need to reinstall Yunohost entirely, or what?
Did you load the yunohost iso or did you install it over Debian 11?
[04:45:13]
<Monitor Kernel Access> Yes,it matters
[04:45:30]
<nullitope> Installed over Debian 11.
[04:45:46]
<Monitor Kernel Access> You might need to resize the ext4 partition
[04:46:28]
<Monitor Kernel Access> If you look in your Stallion settings you should be able to mount a Gparted iso,and using the stallion web console you can actually use the GUI
[04:48:07]
<Monitor Kernel Access> From there boot the iso like any other PC and resize the partition using Gparted
[05:00:35]
<nullitope> Awesome, I think I got it :) Thank you!
[05:25:57]
<Monitor Kernel Access> Glad to help! My DMs are always open if you need more help,fellow Yunohost/BuyVM user!
[07:33:19]
<Canada Goose> I've been noticing this error come up that prevents the php7.4-fpm service from starting correctly every time I reboot my server. I've just been starting the service manually each time after a reboot and then it runs fine until the next reboot. Is this something anyone else has had happen?
-- Boot 4bdde0e8a46246f5b7fe5902a1b3436e --
Nov 13 20:05:39 systemd[1]: Starting The PHP 7.4 FastCGI Process Manager...
Nov 13 20:05:39 php-fpm7.4[542]: [13-Nov-2023 20:05:39] ERROR: [pool synapse] cannot get uid for user 'synapse'
Nov 13 20:05:39 php-fpm7.4[542]: [13-Nov-2023 20:05:39] ERROR: FPM initialization failed
Nov 13 20:05:39 systemd[1]: php7.4-fpm.service: Main process exited, code=exited, status=78/CONFIG
Nov 13 20:05:39 systemd[1]: php7.4-fpm.service: Failed with result 'exit-code'.
Nov 13 20:05:39 systemd[1]: Failed to start The PHP 7.4 FastCGI Process Manager.
[08:02:50]
<tituspijean> Canada Goose: what's the output of `id synapse` ?
[08:07:11]
<Canada Goose> > <@titus:pijean.ovh> Canada Goose: what's the output of `id synapse` ?
ngill@nortel:~$ id synapse
uid=48546(synapse) gid=48546(synapse) groups=48546(synapse),4002(all_users),5001(mail.main),5002(xmpp.main),86966(forgejo.main),6103(synapse.main),13783(cinny.main),20573(nextcloud.main),38138(nextcloud.api),98435(peertube.main),67467(reverseproxy.main),20339(wordpress.main)
[08:08:22]
<tituspijean> weird... php-fpm cannot start due to not finding this uid, but we can see it does exist 😕
[08:08:55]
<Canada Goose> It only seems to fail to start immediately after a reboot too, which I find odd.
[08:09:18]
<Canada Goose> I just restart it manually each time and then it's fine until the next time I reboot
[09:36:44]
<thatoo> Hello, I’m trying to use the Gestion app in Nextcloud, https://apps.nextcloud.com/apps/gestion 4, without success.
I opened a message on the forum and an issue on github to track this issue :
- https://forum.yunohost.org/t/nextclouds-gestion-app-is-not-working/27010
- https://github.com/baimard/gestion/issues/360
I think it is linked to nginx reverse proxy but I don't know what test I could make to find the solution. I'm afraid that if I face this issue with this nextcloud app, it might happen for others.
[09:49:59]
<lapineige> Hedgedoc upgrade fails. Any idea of what's going on ?
https://paste.yunohost.org/raw/daheloguvu
I see no error…
[09:50:28]
<lapineige> Oh wait I didn't see this : `meta-marked@https://github.com/hedgedoc/meta-marked.git#commit=0cb5065253ab0f9851baec34dc9edbb40eabf3d0: The remote archive doesn't match the expected checksum`
[10:02:57]
<thatoo> After reading these two pages :
* [ 405 Method Not Allowed - HTTP | MDN](<https://developer.mozilla.org/fr/docs/Web/HTTP/Status/405>)
* [Résoudre les problèmes d’applications WEB API2 qui fonctionnent dans Visual Studio et échouent sur un serveur IIS de production | Microsoft Learn](<https://learn.microsoft.com/fr-fr/aspnet/web-api/overview/testing-and-debugging/troubleshooting-http-405-errors-after-publishing-web-api-applications>)
it looks like PROPFIND and PROPPATCH , which are two methods used with WebDAV, and they are used to query or set properties for a URI, are not yet handle by yunohost nextcloud nginx conf.
Any idea what I should do to try to make it working?
[10:12:40]
<tituspijean> thatoo: Edit `/etc/nginx/conf.d/<nextcloud domain>.d/nextcloud.conf` to have on line 138 ` if ($request_method ~ ^(PUT|DELETE|PATCH|PROPFIND|PROPPATCH)$) {`
[10:12:41]
<tituspijean> then test and reload nginx: `sudo nginx -t` and `sudo systemctl reload nginx`
[10:14:17]
<tituspijean> btw it's always better to report issues with YunoHost apps to YunoHost and not directly to the upstream app. There is high probability issues lie with YunoHost and not the app itself 😉
[10:17:57]
<thatoo> You rock! It works!
[10:18:19]
<thatoo> I'll report to nextcloud's yunohost app indeed, thanks
[10:35:05]
<tituspijean> > <@titus:pijean.ovh> weird... php-fpm cannot start due to not finding this uid, but we can see it does exist 😕
Canada Goose: ok, I found the issue. The `synapse` user has been create not as a system user, but as a YunoHost user. Its existence is managed by LDAP. If PHP-FPM tries to start before the slapd service, then the user does not "exist" yet.
I don't quite like it, but try to alter the php7.4-fpm service to have it start after slapd:
[10:41:06]
<tituspijean> `sudo systemctl edit php7.4-fpm`
[10:42:47]
<tituspijean> and paste this on line 3:
```
[Unit]
Description=The PHP 7.4 FastCGI Process Manager
Documentation=man:php-fpm7.4(8)
After=network.target
After=slapd.service
[Service]
Type=notify
ExecStart=/usr/sbin/php-fpm7.4 --nodaemonize --fpm-config /etc/php/7.4/fpm/php-fpm.conf
ExecStartPost=-/usr/lib/php/php-fpm-socket-helper install /run/php/php-fpm.sock /etc/php/7.4/fpm/pool.d/www.conf 74
ExecStopPost=-/usr/lib/php/php-fpm-socket-helper remove /run/php/php-fpm.sock /etc/php/7.4/fpm/pool.d/www.conf 74
ExecReload=/bin/kill -USR2 $MAINPID
[Install]
WantedBy=multi-user.target
```
[10:43:02]
<tituspijean> Save and restart php7.4-fpm
[10:43:15]
<tituspijean> Hopefully it fixes the issue on next reboot
[14:47:20]
<TOGLK> Their is a link to a defunct domain when you click the "Demo" button on the following webpage:
https://apps.yunohost.org/app/librex
[14:49:12]
<TOGLK> It links to this domain, which is for sale:
https://librex.beparanoid.de/
[14:50:15]
<tituspijean> Feel free to suggest a fix on https://github.com/YunoHost-Apps/librex_ynh/blob/master/manifest.toml#L14
[14:50:53]
<TOGLK> I'll login and do that.
[14:57:09]
<TOGLK> Should I suggest complete removal or just an alternative domain that's using LibreX?
I don't know of any safe domains to recommend as I have never used LibreX before.
I just came by to point out that it was a broken link.
This is my first time using the YunoHost website.
[14:57:22]
<TOGLK> Just trying to be nice.
[14:58:00]
<TOGLK> And potentially save other people from landing on a domain that gets bought by a malware pusher or other shady people.
[15:06:56]
<Mateusz Szymański> > <@toglk:matrix.org> Should I suggest complete removal or just an alternative domain that's using LibreX?
>
> I don't know of any safe domains to recommend as I have never used LibreX before.
>
> I just came by to point out that it was a broken link.
>
> This is my first time using the YunoHost website.
typically if available apps will offer demo link, so any other public instance should be OK
[15:31:43]
<TOGLK> https://github.com/YunoHost-Apps/librex_ynh/issues/18
[15:52:00]
<Salamandar> I have an hypothetical question
[15:56:39]
<Salamandar> Why are SQL servers on a "one server for every app" paradigm ?
[15:57:02]
<Salamandar> Why aren’t they on a "one server per database" system ?
[15:57:11]
<Salamandar> (NOT in yunohost, just in general)
[15:58:26]
<tituspijean> thanks for the issue! (a Pull Request, or direct file edit would have been perfect, but still thank you very much in helping improve the UX of yunohost 🙂 )
[15:59:04]
<Aleks (he/him/il/lui)> Salamandar: i'm not sure they are "in general", i think a lot of people nowadays use docker / docker-compose paradigm with one DB server per app ?
[16:00:40]
<Salamandar> > <@Alekswag:matrix.org> Salamandar: i'm not sure they are "in general", i think a lot of people nowadays use docker / docker-compose paradigm with one DB server per app ?
Well yeah and no, very often people have
* one docker with mysql
* one docker for nextcloud
* one docker for XX
* one docker for YY
[16:01:20]
<Salamandar> > when this was designed we were not yet in the era of "if the server lacks ressource then just buy more resources"
Yes indeed
[16:01:24]
<Aleks (he/him/il/lui)> but certainly having one server for multiple apps seems more efficient ressource-wise, so probably why it was designed this way (and also relates, I guess, the SQL user system which i always found hella confusing), and probably when this was designed we were not yet in the era of "if the server lacks ressource then just buy more resources"
[16:01:30]
<tituspijean> they are also very good at handling concurrent instructions across multiple databases, and very good at restraining access to these databases.
Adding more servers would add more overhead with little improvement.
[16:01:41]
<Salamandar> About resources, I can’t see much more than RAM usage of loading executables and shared libraries
[16:01:51]
<Salamandar> > at handling concurrent instructions across multiple databases
What do you mean by that ?
[16:02:27]
<Salamandar> Ah you mean they already are good at separating databases
[16:02:52]
<Salamandar> I just see that from a "management" point of view, where having multiple servers are just easier to handle
[16:03:31]
<Aleks (he/him/il/lui)> > <@Salamandar:matrix.org> About resources, I can’t see much more than RAM usage of loading executables and shared libraries
at some point we almost went for "One PHP-FPM process per PHP app" (we somewhat decided that, but never actually implement it) and at that time, we ran a couple experiment and found the master process is like 50 MB-ish ... so if you have, say 10 PHP apps, that's 500 MB or RAM used instead of 50 MB ... i would expect similar result for DB servers
[16:03:53]
<tituspijean> > <@Salamandar:matrix.org> I just see that from a "management" point of view, where having multiple servers are just easier to handle
I'm not sure about that, for example we install PostgreSQL or MariaDB from their debian repos, and I think running multiple of them is not straightforward
[16:04:05]
<Salamandar> > <@Alekswag:matrix.org> at some point we almost went for "One PHP-FPM process per PHP app" (we somewhat decided that, but never actually implement it) and at that time, we ran a couple experiment and found the master process is like 50 MB-ish ... so if you have, say 10 PHP apps, that's 500 MB or RAM used instead of 50 MB ... i would expect similar result for DB servers
ouch…
[16:04:07]
<Aleks (he/him/il/lui)> 500 MB of RAM is nothing these days tho, but for SBC that's the difference between being able to run 10 apps on a RPi ~2, and not being able to
[16:04:07]
<Salamandar> alright
[16:04:08]
<tituspijean> (or simply with multiple services? I actually don't know)
[16:04:21]
<Salamandar> > <@Alekswag:matrix.org> 500 MB of RAM is nothing these days tho, but for SBC that's the difference between being able to run 10 apps on a RPi ~2, and not being able to
yeah my Helios64 is still 4GB of ram…
[16:04:24]
<Salamandar> zfs is sad
[16:04:44]
<Aleks (he/him/il/lui)> well 4GB of RAM should be considered "huge"
[16:05:07]
<Salamandar> eeeh
[16:05:12]
<Salamandar> ¯\_(ツ)_/¯
[16:05:16]
<tituspijean> > <@Alekswag:matrix.org> at some point we almost went for "One PHP-FPM process per PHP app" (we somewhat decided that, but never actually implement it) and at that time, we ran a couple experiment and found the master process is like 50 MB-ish ... so if you have, say 10 PHP apps, that's 500 MB or RAM used instead of 50 MB ... i would expect similar result for DB servers
erf, I almost put that on the table again today, regarding PHP7.4-FPM for synapse trying to run before SLAPD
[16:06:02]
<tituspijean> (but the question is rather why the f* Synapse needs a "human" user in LDAP)
[16:06:27]
<Aleks (he/him/il/lui)> > <@Alekswag:matrix.org> well 4GB of RAM should be considered "huge"
let me rephrase it :
4GB of RAM should be considered huge, considering how much "modern" development tends to waste resources comparatively to what we achieve, compared to what people were able to achieve in the past with, say, 32kb of RAM
[16:07:26]
<Aleks (he/him/il/lui)> i plaid guilty, i develop in Python and am happy to have the vast majority of variables "stringly-typed"
[16:08:16]
<eric_G> we used to send people to the moon with less than 32 KB of RAM 😶🌫️
[16:08:25]
<Aleks (he/him/il/lui)> > <@titus:pijean.ovh> (but the question is rather why the f* Synapse needs a "human" user in LDAP)
yeah that's a real question, i didn't really dig the topic deep but that sounds unecessary ...
[16:08:39]
<eric_G> just me contributing to the discussion😅
[16:09:10]
<Aleks (he/him/il/lui)> > <@ericg:matrix.org> we used to send people to the moon with less than 32 KB of RAM 😶🌫️
yeah ... if we can go to the room with 32kb of RAM, surely we can probably find a way to build mailing list and IM systems that do not require 16 GB of RAM idk
[16:10:15]
<Aleks (he/him/il/lui)> but a good chunk of it is the mindset and availability of resource, if throwing additional GBs of RAM at the server is super cheap and saves development time, then surely ...
[16:10:26]
<tituspijean> > <@Alekswag:matrix.org> yeah that's a real question, i didn't really dig the topic deep but that sounds unecessary ...
looks like it needs to authenticate in the LDAP: https://github.com/YunoHost-Apps/synapse_ynh/blob/9eaddbe4123ebc2a6774f664dc525adeacb3de57/conf/homeserver.yaml#L2250
[16:10:56]
<Aleks (he/him/il/lui)> yes but why does it needs to authenticate to LDAP ...
[16:11:00]
<Aleks (he/him/il/lui)> plenty of apps do anonymous queries
[16:13:01]
<tituspijean> > if we can go to the room with 32kb of RAM
that's definitely why I am affected by the doorway effect, forgetting what I was meant to do when going into a room 😛
[17:13:56]
<lapineige> > <@Alekswag:matrix.org> let me rephrase it :
>
> 4GB of RAM should be considered huge, considering how much "modern" development tends to waste resources comparatively to what we achieve, compared to what people were able to achieve in the past with, say, 32kb of RAM
and considering how much it cost and the environnemental cost too
[17:15:03]
<lapineige> > Hedgedoc upgrade fails. Any idea of what's going on ?
> https://paste.yunohost.org/raw/daheloguvu
> I see no error…
ping eric_G btw as you might have seen this in the past ?
[17:15:54]
<Aleks (he/him/il/lui)> npm singing the song of its people
[17:18:11]
<Aleks (he/him/il/lui)> if your dependencies can't be found in the cache and will be fetched from the remote registry, clap your hand
[17:30:59]
<lapineige> but there's a shasum issue
[20:18:24]
<Canada Goose> > <@titus:pijean.ovh> and paste this on line 3:
>
> ```
> [Unit]
> Description=The PHP 7.4 FastCGI Process Manager
> Documentation=man:php-fpm7.4(8)
> After=network.target
> After=slapd.service
>
> [Service]
> Type=notify
> ExecStart=/usr/sbin/php-fpm7.4 --nodaemonize --fpm-config /etc/php/7.4/fpm/php-fpm.conf
> ExecStartPost=-/usr/lib/php/php-fpm-socket-helper install /run/php/php-fpm.sock /etc/php/7.4/fpm/pool.d/www.conf 74
> ExecStopPost=-/usr/lib/php/php-fpm-socket-helper remove /run/php/php-fpm.sock /etc/php/7.4/fpm/pool.d/www.conf 74
> ExecReload=/bin/kill -USR2 $MAINPID
>
> [Install]
> WantedBy=multi-user.target
> ```
Either I'm doing something wrong or this didn't work
[20:18:25]
<Canada Goose> I got the same error after adding `After=slapd.service`
[23:59:48]
<Aleks (he/him/il/lui)> Did you run systemctl daemon-reload