Bug 278592

Summary:	Phabricator is out of support
Product:	Services	Reporter:	Mirek Chocholous <mirek>
Component:	Code Review	Assignee:	Phabric Admin <phabric-admin>
Status:	Closed Not A Bug
Severity:	Affects Some People	CC:	alster, bofh, vvd
Priority:	---
Version:	unspecified
Hardware:	Any
OS:	Any

Description Mirek Chocholous 2024-04-26 14:53:56 UTC

The Phabricator tool is out of support for some time (officially since June 2021).
https://github.com/phacility/phabricator/commit/9ceb66453501d461800617b9fb2d4af4ba2c9514

I know it's a great tool with not so many bugs but it might be wise to update / migrate, already.


Easy / doable should be to migrate to Phorge - a community driven fork of Phabricator and it would keep all the current flows working.
https://we.phorge.it/w/installation_and_setup/update_from_phabricator/


Other option is to migrate to something completely different or keep Phabricator until it will keep running.

Comment 1 Vladimir Druzenko freebsd_committer

2024-04-26 14:58:48 UTC

*** Bug 278593 has been marked as a duplicate of this bug. ***

Comment 2 Muhammad Moinur Rahman freebsd_committer

2024-04-26 15:50:08 UTC

First of all thanks for the notification. Here are my takes on the ticket:

1. We are very much aware about the situation.
2. We do not run the stock Phabricator and we have some customization and patches. Migrating to Phorge is not an easy option. We have to slipstream these patches in a forked version. And also regularly maintain that fork as these patches are specific to our situation and should not go to upstream.
3. I am mostly in charge of day to day maintenance of our Phabricator and the patches are not easy to be slipstreamed and it requires time. Unfortunately none has that free time to work that long hour.
4. Currently cluster access is extremely limited to selected committers only. And after the incidents with xz I believe more precautions will be taken into account. And even beyond committers unless a committer has a proven track record of being trustworthy no access will be provided to the cluster.
5. We are working on a single platform tool for everything and mostly various applications like Gitea, Forgejo, Gitlab are being tested for these cases and we will most probably migrate to one of those. So I find it more lucrative to work on those rather than going to fix Phabricator at the moment.
6. Maintenance is a burden for our Phabricator instance. Current DB size is somewhat more than half a terra and the entire data folder is also more than 2 terras.

So I do not see this as a bug itself as things are rolling it's just not in the pace it should be. If you have a concern or something to offer to help please join the IRC channel in EFNet where there is a #gitcvt channel. If you have something else to add on this regard please send a mail to phabric-admin@.

Comment 3 Älven 2024-04-27 13:00:26 UTC

Hello!

May hosted SourceHut instance be added to consideration, please?
It is modular and has numerous advantages to offer,

* Hosted both git and mercurial repositories
* Powerful continuous integration (with *BSD virtualised builds!)
* Mailing lists & code review tools (full support for email-based workflow!)
* Focused ticket tracking
* Hosted real-time chat services
* Sophisticated account management & security
* Markdown- and git-driven wikis
* Pastebin and static web pages hosting
* and many more

I believe it could be the best all-in-one solution for FreeBSD project, and may hope you will like it too :)

https://sourcehut.org/

Comment 4 Muhammad Moinur Rahman freebsd_committer

2024-04-27 13:51:39 UTC

(In reply to Andrey Korobkov from comment #3)
Hi,

Sorry but this is a no go for various other reasons. This has had been in our consideration but we no longer consider this as an option.

Comment 5 Älven 2024-04-27 14:19:38 UTC

Sad to hear this. :( I hoped for the best, but… sad.

May at least I know the reasons, why it is not an option anymore?
Maybe some link to some topic where it was discussed, please?

Comment 6 Muhammad Moinur Rahman freebsd_committer

2024-04-27 14:29:38 UTC

(In reply to Andrey Korobkov from comment #5)
There were various reasons but so far I recall it was due to LDAP integration and it's development model. These are mainly discussed in the gitcvt meeting and you can join the #gitcvt channel in EFNet and discuss there.