[05:53:04]
<tituspijean> It looks like the ACL are not supported on the drive holding the multimedia directory. Do you have mounted it from outside an Ext4 (or other "linux" filesystem)?
[06:02:59]
<Guillaume Bouzige> hello, I am trying to update the facilmap app but I run into issue such as 'FATAL ERROR: Reached heap limit Allocation failed - JavaScript heap out of memory'
[06:03:20]
<Guillaume Bouzige> I have already added 2G of swap since my available RAM is only 1G
[06:03:55]
<Guillaume Bouzige> how can I get the upgrade to execute ? do I need more swap or is there a way ?
[06:25:36]
<furyventura> > <@titus:pijean.ovh> It looks like the ACL are not supported on the drive holding the multimedia directory. Do you have mounted it from outside an Ext4 (or other "linux" filesystem)?
Yes, it is an NFS mount. So my NAS needs the setfacl installed as it is missing? I will give it a try. Is there any possibility to run the user creation for the existing users or do I have to delete and recreate?
[06:28:31]
<furyventura> > <@botagiuks:tiesiog.lt> I recently had some experience with the `setfacl` command when I was setting up a shared directory between two containers, one of which was yunohost with Sonarr/Radarr, and one was a downloader client.
> Essentially, with setfacl you can set default permissions in a directory that are set on top of the permissions that the user that creates the files decides to set them to. At least that's what I needed it for, because when a different container creates / moves files to a shared storage, the other container most likely will have difficulties accessing / grabbing / etc. the files, which is why having `setfacl` helps.
> Now to answer your question, I would speculate that the command here wants to add default permissions for the multimedia directory and its contents, but presumably fails.
> Each user has separate multimedia folders and although I do not know why that is, you can verify whether the permissions are incorrect by using `getfacl /home/yunohost.multimedia/<user>/`, which will tell you the full permission list for that directory.
Thank you very much. I will check and see how to fix this. Now I can check the internet as I understand better the cause for the problem
[06:36:34]
<tituspijean> > how can I get the upgrade to execute ? do I need more swap or is there a way ?
with the current state of packaging and your hardware, adding swap seems to be the only way
[06:36:42]
<Guillaume Bouzige> shall I add more
[06:37:22]
<Guillaume Bouzige> also the warning message seem to not take in account the swap memory available
[06:39:09]
<Guillaume Bouzige> indeed but how to be sure it is used by my java heap memory thing
[07:18:36]
<tituspijean> I chose 2800 MiB as a close value to 3GB of RAM, but maybe you should fine-tune it to your current RAM usage
[07:18:36]
<tituspijean> > indeed but how to be sure it is used by my java heap memory thing
I see, maybe try running the upgrade from command line after running first `export NODE_OPTIONS="--max-old-space-size=2800"` 🤞 (no idea if the env variable will be propagated to the node helper, otherwise you may need to hack into it, and that would be a great addition to it)
[07:18:36]
<tituspijean> > <@furyventura:matrix.org> Yes, it is an NFS mount. So my NAS needs the setfacl installed as it is missing? I will give it a try. Is there any possibility to run the user creation for the existing users or do I have to delete and recreate?
I'll let you investigate if whether or not NFS supports ACLs made by a remote server.
In the mean time, the warnings are triggered by a post_user_create hook: https://github.com/YunoHost/yunohost/blob/c6bda180b4e83d612e37340f7a3e96cbce98b0a2/hooks/post_user_create/ynh_multimedia#L28
You can comment out the `setfacl` lines so that YunoHost does not try to set them up. The file is in `/usr/share/yunohost/hooks/post_user_create/ynh_multimedia`.
[07:18:36]
<tituspijean> > <@titus:pijean.ovh> I'll let you investigate if whether or not NFS supports ACLs made by a remote server.
>
> In the mean time, the warnings are triggered by a post_user_create hook: https://github.com/YunoHost/yunohost/blob/c6bda180b4e83d612e37340f7a3e96cbce98b0a2/hooks/post_user_create/ynh_multimedia#L28
> You can comment out the `setfacl` lines so that YunoHost does not try to set them up. The file is in `/usr/share/yunohost/hooks/post_user_create/ynh_multimedia`.
Bear in mind it will get overwritten upon next YunoHost upgrade. YunoHost needs to add a test to check that ACL are supported on the filesystem before running the commands.
[07:18:37]
<orhtej2> > <@titus:pijean.ovh> I see, maybe try running the upgrade from command line after running first `export NODE_OPTIONS="--max-old-space-size=2800"` 🤞 (no idea if the env variable will be propagated to the node helper, otherwise you may need to hack into it, and that would be a great addition to it)
it will likely not, tried that in the past :/
[07:18:37]
<Guillaume Bouzige> > it will likely not, tried that in the past :/
what do you mean ? I have changed the swap to 5GB see if it pass.
[07:20:42]
<Guillaume Bouzige> failed the same
[07:20:56]
<Guillaume Bouzige> I kind of feel it that it doesnt get the swap...
[07:21:43]
<Guillaume Bouzige> does this `export NODE_OPTIONS="--max-old-space-size=2800"` works ?
[07:46:40]
<orhtej2> > does this `export NODE_OPTIONS="--max-old-space-size=2800"` works ?
It needs to get added at callsite
[07:46:41]
<orhtej2> With `sudo` ofc
[07:46:41]
<orhtej2> > does this `export NODE_OPTIONS="--max-old-space-size=2800"` works ?
`yunohost app upgrade facilmap -u https://github.com/orhtej2/facilmap_ynh/tree/heap`
[07:46:41]
<orhtej2> expect this to take ages as swap is slow, to the point where I'd run inside `screen`
[07:46:41]
<Guillaume Bouzige> I see
[07:57:29]
<Guillaume Bouzige> failed
[08:04:12]
<Guillaume Bouzige> yup i think so but the screen session got closed
[08:04:12]
<tituspijean> same error?
[09:04:50]
<tituspijean> @guillaume_bouzige:matrix.org it has most likely crashed. If you need to reboot your server, before running again the upgrade try to check how much memory is actually free on your system. 2800 may be too high
[09:04:50]
<orhtej2> Also log should be available in webadmin
[09:04:50]
<orhtej2> Screen -r
[09:04:51]
<Guillaume Bouzige> after reboot I have 1.1G available as RAM and 3G as Swap
[09:09:29]
<Guillaume Bouzige> same error with orhtej2 repo
[09:09:41]
<Guillaume Bouzige> `FATAL ERROR: Reached heap limit Allocation failed - JavaScript heap out of memory`
[09:11:22]
<Aleks (he/him/il/lui)> :|
[09:11:27]
<Aleks (he/him/il/lui)> *javascript*
[09:11:37]
<Guillaume Bouzige> yeah nodejs
[09:12:07]
<Guillaume Bouzige> ` DEBUG - [facilmap-frontend]: FATAL ERROR: Reached heap limit Allocation failed - JavaScript heap out of memory` frontend compil
[09:12:28]
<Aleks (he/him/il/lui)> our ancestors are probably crying knowing how we waste computer resources despite having supercomputers in our pockets
[09:12:40]
<Guillaume Bouzige> ahahah indeed
[09:16:14]
<Guillaume Bouzige> so that export NODE OPTION shall be placed on the frontend compilation no ?
[09:18:04]
<orhtej2> > so that export NODE OPTION shall be placed on the frontend compilation no ?
It is, can you share full log? I patched just one invocation
[09:18:04]
<Guillaume Bouzige> https://paste.yunohost.org/raw/rupibiguwo
[09:19:37]
<banthex> hihi. are there any guides how to update apps by hand? i need a photoprism update and im searching for a manuell^^
[09:21:42]
<orhtej2> > <@banthex:matrix.org> hihi. are there any guides how to update apps by hand? i need a photoprism update and im searching for a manuell^^
wdym? From catalog or different upstream?
[09:21:42]
<Aleks (he/him/il/lui)> (i guess they mean bypassing yunohost because they want the app to be upgraded somehow)
[09:24:16]
<orhtej2> > https://paste.yunohost.org/raw/rupibiguwo
Clearly my patch did not work as it tries to use 1gb of ram. Perhaps vite is different in that regard? Im out of ideas
[09:24:17]
<Guillaume Bouzige> I dunno neither but I add similar issues in the past compiling on small hardware big nodejs projects...it is hard
[09:24:17]
<banthex> > wdym? From catalog or different upstream?
i guess a diffrent upstream. the latest version of yunohost hast no 2 factor auth. the docker version has
[09:35:59]
<orhtej2> Or you can change the sources and hope for the best
[09:35:59]
<orhtej2> > <@banthex:matrix.org> i guess a diffrent upstream. the latest version of yunohost hast no 2 factor auth. the docker version has
I believe you can install docker app and expose it via redirect_ynh but never tried that
[09:35:59]
<Aleks (he/him/il/lui)> (there are 4 opened PRs : https://github.com/YunoHost-Apps/photoprism_ynh/pulls )
[09:36:24]
<banthex> > I believe you can install docker app and expose it via redirect_ynh but never tried that
Yes thanks i wanted to avoid that. Big thanks!
[11:20:53]
<bharathmp> So I checked with a handful ISPs in my region and none of them allow us to send through port 25, so I guess using a relay is the only option. Are there any free relays that any of you would recommend? My use case is pretty much to allow user registration (2 users) on the home-servers for Yunohost applications that require email based sign-ups (such as Vaultwarden, etc.). I won't be using it as a mail service in itself. The yunohost server is running on a Raspberry Pi and will continue to do so for the foreseeable future.
[12:44:57]
<@chrichri:ween.de> > <@bharathmp:matrix.org> So I checked with a handful ISPs in my region and none of them allow us to send through port 25, so I guess using a relay is the only option. Are there any free relays that any of you would recommend? My use case is pretty much to allow user registration (2 users) on the home-servers for Yunohost applications that require email based sign-ups (such as Vaultwarden, etc.). I won't be using it as a mail service in itself. The yunohost server is running on a Raspberry Pi and will continue to do so for the foreseeable future.
Why don't you just enable email and use the local accounts via imap to interact with your local services? Having another imap-account for two users doesn't seem that much hassle compared to running over a relay...
[18:05:56]
<bharathmp> > <@chrichri:ween.de> Why don't you just enable email and use the local accounts via imap to interact with your local services? Having another imap-account for two users doesn't seem that much hassle compared to running over a relay...
I do have email enabled. But I continue to get the warning the port 25 is not open. And the emails sent my Vaultwarden are stuck in the cue. They don’t get delivered at all. Could you point me to some resource that would help me configure it in this manner?
[18:27:21]
<@chrichri:ween.de> > <@bharathmp:matrix.org> I do have email enabled. But I continue to get the warning the port 25 is not open. And the emails sent my Vaultwarden are stuck in the cue. They don’t get delivered at all. Could you point me to some resource that would help me configure it in this manner?
I don't know vaultwarden. The integration I made uses tcp:587 on localhost with a user and a password. The [integration of vaultwarden](https://github.com/YunoHost-Apps/vaultwarden_ynh/blob/master/manifest.toml) seems to offer the sending of email by its system user (probably 'vautlwarden'). I guess that the from field needs to be set to the email <system user>@<your_host.doma.in> for this to work. Probably there's an entry in /etc/postfix/app_senders_login_maps for vaultwarden. If that entry exists it wouldn't be a good idea to have an alias or yunohost user with the same email address: https://github.com/YunoHost/yunohost/pull/1843 .
Finally I'd say the question needs to be answered by the maintainers of the integration. You can contribute to the issues [here](https://github.com/YunoHost-Apps/vaultwarden_ynh/issues/262).
[20:09:49]
<Paprika> I am not very well versed in that manner, but shouldn’t mails within the local network go through regardless?
As long as you have a static DNS entry for your domain / IP, I’m pretty sure you’re good to go for local delivery?