Alternative to manually install 3rd party repositories

There is an alternative to manually manage repositories and keys and that is to use extrepo

extrepo is a curated list of 3rd party repositories and keys and it's a debian package.
It's only been around a couple of years so I don't know how widely used it is yet.

Installation

To install it run

apt install extrepo Add repository

To add postgreSQL repository for example:

extrepo enable postgresql Disable repository

To disable a repository, for example:

extrepo disable postgresql Where do files go?

extrepo puts apt config files in /etc/apt/sources.list.d as you would manually but manages keys in it's own directory /var/lib/extrepo/keys

Repositories available

Currently these repositories are in there:

anydesk apertium-nightly apertium-release bareos belgium_eid_continuous brave_beta brave_nightly brave_release caddyserver consol debian_official dns-oarc docker-ce edge elbe eturnal eyrie fai feistermops gitlab_ce gitlab_ee gitlab_runner google_chrome google_cloud grafana grafana_beta grafana_enterprise grafana_enterprise_beta haproxy-2.8 i2pd janitor jellyfin jenkins jitsi-stable kea keybase kicksecure kicksecure_developers kicksecure_proposed kicksecure_testers lihas liquorix matrix mobian msteams neurodebian_software newrelic nginx node_12.x node_14.x node_16.x node_18.x notesalexp ooni openmodelica-contrib-nightly openmodelica-contrib-release openmodelica-contrib-stable openmodelica-nightly openmodelica-release openmodelica-stable openstack_antelope openstack_zed openvpn opera_stable opsi passbolt postgresql prosody proxmox-ceph-quincy proxmox-pve proxmox-pve8 r-project raspberrypi raspbian-addons realsense rspamd signal skype slack speedtest-cli spotify steam surface-linux sury syncevolution syncthing teamviewer_default teamviewer_preview torproject trinity vector vscode vscodium weechat whonix whonix_developers whonix_proposed whonix_testers winehq wire-desktop wire-internal-desktop wtf wtf-lts x2go x2go-extras x2go-lts x2go-nightly xpra xpra-beta yarnpkg zammad zulu-openjdk

@JaredBusch said in Zerotier installs but no conectivity on Fedora 38:

@syko24 said in Zerotier installs but no conectivity on Fedora 38:

@JaredBusch - I know their site says Windows, but maybe something related with the firewall rules getting messed up.

https://discuss.zerotier.com/t/attention-zerotier-on-windows-users-please-update-your-client-s-to-1-10-6-or-later/12706

I temporarily shutdown firewalld on both systems as well as disabled selinux.

While I doubt it's the issue - ZT has it's own firewall rules, any possible issue there?

@popester said in Ubiquity U-LTE:

My brand new U-LTE cellular failover device from Ubiquity appears to have lost its APN setting. I guess default is not the Dallas Fort Worth metro plex in texas. It is the one that AT&T has its hooks into, or at least from what i understand. Has this happened to someone else? I can get into the config but do not want to touch any settings for fear of doing more harm than good. I reached out to Ubiquity Support and they shot back that I needed to get the correct settings for the device from AT&T. Not real sure how to do that. Have found lots of information. Since i am paying ubiquity a monthly charge for the service I am not sure who to reach out to. Anyway. Any help would be appreciated.

Have you had any contact with AT&T, or was it all Ubiquity?

I deal with cellular data things as the majority of my work now, so I may be able to help.

oh n/m you were upgrading.. yeah. you just need to sed all of the repo.list you are using. Each system could be different depending on how you added the no sub repo.

Some instructions have that being added as a separate repo.list file.

@scottalanmiller said in Installing Mastodon 4.1.2 on Debian 11:

@stacksofplates said in Installing Mastodon 4.1.2 on Debian 11:

Their official image is here ghcr.io/mastodon/mastodon

Yeah, that one was flaky. But is it official? Why's it on a different account?

Yes it's their official. Ghcr is the GitHub packages repo. It's similar to GCR, ECR, etc. Other OCI image repositories are more popular now since Docker has imposed limits on Docker Hub.

Leaving this new system standalone, not joining the old cluster or creating a single system cluster.

The new (added in 7.3) qm remote-migrate works well. The docs suck, but the function is great.
bb4d3436-413c-4121-bbba-92032927da78-image.png

3124c7d3-2c98-4f08-966d-5d2089534292-image.png

9247c618-4860-410c-880d-b53698015d98-image.png

@Eric-Ross said in CentOS 7 VM won't boot after migration to Proxmox:

@JaredBusch Ah, too bad that wasn't the fix.

I didn't bother at the dracut point. I likely could have recovered the system. I simply decided to stop putting off the migration of those workloads.

They were still running CentOS 7 after all.

@scottalanmiller said in Configure Mastodon to Use Zoho ZeptoMail for SMTP Email:

@PhlipElder said in Configure Mastodon to Use Zoho ZeptoMail for SMTP Email:

All of our mail servers are properly set up with a SmartHost, SPF, and DMARC. We don't have any reputation issues. That's a 2000s kind of problem.

No, it's current. Microsoft blacklists entire datacenters still. There's a lawsuit in Illinois about it that they lost. It's a very real thing still that RBL are used by the majors (like Microsoft, but not Google) and used to shut down entire ISPs and IP blocks. I've been brought in to help providers with these issues, it's a huge legal problem and if you don't get hit with it, it seems like it doesn't exist, and if you do get hit, it seems like there's no recourse (hint: there isn't.)

We're dealing with a client who has their site hosted in a Yandex.RU data centre and guess what? Yeah, some putz between their web server and ProofPoint has a sinkhole because Ukraine. The web server can e-mail @Outlook.Com because they have servers around the world but not ProofPoint.

We've had lots of issues over the years where one midbone/backbone provider either shapes or blocks packets from another because they're in a contract dispute. Poof. Packets gone.

oops - remotedesktop.google.com/access is the correct url - mistyped

@DustinB3403 said in Windows Server Licensing:

HPE has some tool which indicates the inverse here

This tool is correct, as far as I've last known.

@IgnaceQ said in If starting from scratch, would you suggest learning BASH or Ansible?:

Besides anisble there are other tools for remote tooling. In a linux world you need bash. this is a basic need.
On top of bash, it really depends on the environmnet you're in. I would suggest python on top of this. You can use python in ansible, chef, terraform, etc... Also on non linux environments, python is availlable (bsd, windows).

He needs Ansible because he's doing an Ansible based certification path.

@dafyre Yes, I did. I am using a different kernel now and for the moment, things seem to be happy! (Ish!)

I ASSUME that their answer is that no one should be emailing attachments like that and they should be sending links to the hosted files instead?

I can't remember the last time that we had to email someone an attachment of an office document, just saying that sounds like a legacy process. Who needs to do that in the modern world?

@NETSERVER that's a bit surprising, but honestly, I guess not. With the way that IBM decided to view CentOS and RHEL, it isn't the stable, well supported system that it used to be. It used to be taken far more seriously. Now, I can't imagine a scenario where I'd deploy RHEL. And at one time, I had one of, if not the, largest RHEL fleets in the world.

I think what you want is a completely external device. That's how this is normally handled. Meaning it's common to have a dialer / ringer on a computer but you answer the phone. Same thing could be done to make a loudhorn blast anything you want as well.

@CCWTech said in Unable to connect Ubuntu with Google Online Accounts:

Jun 08 11:15:59 CCW-HAL systemd[3523]: vte-spawn-8b4a59a3-a060-4bdd-92e6-285656bdb9a2.scope: Consumed 3.448s CPU time.
Jun 08 11:16:01 CCW-HAL gnome-shell[3803]: g_closure_unref: assertion 'closure->ref_count > 0' failed
Jun 08 11:16:01 CCW-HAL gnome-shell[3803]: g_closure_unref: assertion 'closure->ref_count > 0' failed
Jun 08 11:16:01 CCW-HAL gnome-shell[3803]: g_closure_unref: assertion 'closure->ref_count > 0' failed
Jun 08 11:16:29 CCW-HAL systemd[3523]: Started app-gnome-org.gnome.Terminal-10924.scope - Application launched by gnome-shell.
Jun 08 11:16:29 CCW-HAL dbus-daemon[3552]: [session uid=1000 pid=3552] Activating via systemd: service name='org.gnome.Terminal' unit='gnome-terminal-server.service' requested by ':1.167' (uid=1000 pid=10927 comm="/usr/bin/gnome-terminal.real" label="unconfined")
Jun 08 11:16:29 CCW-HAL systemd[3523]: Starting gnome-terminal-server.service - GNOME Terminal Server...
Jun 08 11:16:29 CCW-HAL dbus-daemon[3552]: [session uid=1000 pid=3552] Successfully activated service 'org.gnome.Terminal'
Jun 08 11:16:29 CCW-HAL systemd[3523]: Started gnome-terminal-server.service - GNOME Terminal Server.
Jun 08 11:16:29 CCW-HAL systemd[3523]: Started vte-spawn-9cee1911-372a-4bb4-8b57-694984e43990.scope - VTE child process 10955 launched by gnome-terminal-server process 10931.
Jun 08 11:16:49 CCW-HAL gnome-control-c[8079]: Error showing account: Child process exited with code 1
Jun 08 11:16:53 CCW-HAL gnome-online-accounts-panel.desktop[9764]: GLib-GIO: Using cross-namespace EXTERNAL authentication (this will deadlock if server is GDBus < 2.73.3)GLib-GIO: _g_io_module_get_default: Found default implementation gvfs (GDaemonVfs) for ‘gio-vfs’GLib: unsetenv() is not thread-safe and should not be used after threads are createdGLib-GIO: _g_io_module_get_default: Found default implementation dconf (DConfSettingsBackend) for ‘gsettings-backend’GoaBackend: Loading all providers: GoaBackend: - googleGoaBackend: - owncloudGoaBackend: - windows_liveGoaBackend: - exchangeGoaBackend: - lastfmGoaBackend: - imap_smtpGoaBackend: - kerberosGoaBackend: activated kerberos providerGLib-GIO: _g_io_module_get_default: Found default implementation gnutls (GTlsBackendGnutls) for ‘gio-tls-backend’Failed to create account: Dialog was dismissed
Jun 08 11:16:53 CCW-HAL xdg-desktop-por[3907]: Realtime error: Could not map pid: Could not determine pid namespace: Could not find instance-id in process's /.flatpak-info
Jun 08 11:17:01 CCW-HAL CRON[11093]: pam_unix(cron:session): session opened for user root(uid=0) by (uid=0)
Jun 08 11:17:01 CCW-HAL CRON[11094]: (root) CMD (cd / && run-parts --report /etc/cron.hourly)
Jun 08 11:17:01 CCW-HAL CRON[11093]: pam_unix(cron:session): session closed for user root
Jun 08 11:17:06 CCW-HAL kernel: WebKitWebProces[11026]: segfault at 55bd22ad9adc ip 000055bd22ad9adc sp 00007ffd1f6adbe8 error 14 likely on CPU 2 (core 2, socket 0)
Jun 08 11:17:06 CCW-HAL kernel: Code: Unable to access opcode bytes at 0x55bd22ad9ab2.
Jun 08 11:17:36 CCW-HAL gnome-shell[3803]: Window manager warning: WM_TRANSIENT_FOR window 0x3a02767 for 0x3a02778 window override-redirect is an override-redirect window and this is not correct according to the standard, so we'll fallback to the first non-override-redirect window 0x3a006dc.
Jun 08 11:17:42 CCW-HAL systemd[1]: Starting systemd-tmpfiles-clean.service - Cleanup of Temporary Directories...
Jun 08 11:17:42 CCW-HAL systemd[1]: systemd-tmpfiles-clean.service: Deactivated successfully.
Jun 08 11:17:42 CCW-HAL systemd[1]: Finished systemd-tmpfiles-clean.service - Cleanup of Temporary Directories.
Jun 08 11:17:42 CCW-HAL systemd[1]: run-credentials-systemd\x2dtmpfiles\x2dclean.service.mount: Deactivated successfully.
Jun 08 11:18:10 CCW-HAL gnome-shell[3803]: Window manager warning: WM_TRANSIENT_FOR window 0x3a029e2 for 0x3a029f0 window override-redirect is an override-redirect window and this is not correct according to the standard, so we'll fallback to the first non-override-redirect window 0x3a006dc.

Looks like it may be a known bug: https://launchpad.net/bugs/2019739

@Pete-S said in Debian 11 & php8:

Not a challenge at all but the reason to run "stable" is for stability.

Once you start abandoning the integration, though, you are abandoning stability. The idea of using an LTS and then replacing the parts of the OS that aren't up to date is counterproductive. Choose the most up to date, best supported, most stable version and use the fully tested and integrated components instead.

The idea of "stable" is not stability in IT terms, that's a myth. It's actually against that. The idea of current is for IT stability. Stable, in reference to an OS like this, is in reference to the versions of products remaining stable so that unsupported, out of date software from bad vendors can be used without updating for long periods of time. Not a positive stable, it's a bad stable.

@WrCombs said in Reboot resets Desktop Win 10 -:

So found out the user is actually corrupt, this is the 4th time this has happened according to the internal IT team, and they're looking to get the PC replaced

It’s always best to replace a corrupt user. The can leave so much of a mess.