- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
All Stories
Fri, Oct 18
What is the normal memory usage for the update daemon during the repository update? It seems to be peaking around 8GB per process for our repositories. Maybe I need to add more RAM, i.e. 32 GB to satisfy this, but that's not easy on AWS.
Repository update is based on how recent the last commit is, but that's only for collecting the list of commits to analyze.
But that should be done in the Daemons, which should at least recover nicely and not break the web side of the application. I don't know if they should be running as php or php-fpm.
Thu, Oct 17
OK, not the monitoring issue.
Wed, Oct 16
@Iniquity - Problem solved by removing that sentence lol - thanks. Now seems super-freshy
Thanks! :)
@BlankEclair Do you want a little help to land this?
@Iniquity - Problem solved by removing that sentence lol - thanks. Now seems super-freshy
Arcanist is misspelled as "Archanist" in the 3rd step of "Updating" section.
Oct 15 2024
Uh! Out of curiosity, what does the notification look like in your web notifications?
Okay, I got it! For future reference if anyone in a similar situation is stumbling across this and wants a fully working example to go off of, my more-than-a-little-cargo-culted code is currently:
Oct 14 2024
Many thanks, @20after4! That all largely did it; in addition to setting up the object a bit wrong (I knew it was a ManiphestTask instance, but was failing to get the actual object of it), I was also just using essentially my:custom-field rather than std:maniphest:my:custom-field.
Appears to be internal monitoring software causing this (not phorge), thanks for the replies.
Oct 10 2024
Noting a user complaint (I'm using a mildly hacked-up Phorge to host https://blog.styx-os.org/ - but the font stack is unchanged): https://fe.disroot.org/objects/5d376f6a-0acc-44fc-bd37-31de8630b647
Oct 9 2024
In T15947#19599, @avivey wrote:Maybe express it as allowed by Policy xxx? (Related to T15277)
So essentially this:
If I were starting today I would probably design the back-end APIs first, then make the web interface derive from those APIs, such that web requests and api requests are not really any different, at least with regard to enforcing access controls.
Maybe express it as allowed by Policy xxx? (Related to T15277)
No clue :)
P.S. maybe relevant - maybe not :D lol https://we.phorge.it/T15945#19568
This one seems worthy of a rubber stamp. And a gold star.
Are you importing some huge git repo or something? I haven't seen a lot of OOM with php-fpm on phorge/phabricator. It's generally not been an issue in the past so this is either a new bug or something specific to the environment. Then again, I've almost always ran Phorge with much more than 8GB of ram available.
For another example, here is a fairly straightforward use of that API which takes place entirely outside of a custom field subclass:
There is a pretty complicated and not really easy to follow subclass of ManiphestCustomField over here: https://phabricator.wikimedia.org/source/phab-extensions/browse/wmf%252Fstable/src/customfields/ReleaseDetailsCustomField.php
@keithzg what you might be missing is
$object = $this->getObject()
There is a significant amount of Phabricator dark matter out there - companies/people using the software, it works well enough, not really easy to know they exist or anything about their usage. I'm sure at least some of them have moved to Phorge. Automattic/wordpress.com have moved to Phorge and I wasn't even aware that they were using Phabricator before that. This is despite the fact that I did a pretty extensive amount of research to identify every company using Phabricator back in ~2019 as part of my work for Wikimedia, with the goal of reaching out and trying to organize an informal Phabricator users group. We had the idea that the various corporate users probably had good reasons to be collaborating and at least talking to each other since most of them were not active in the upstream project. Anyway, that never really panned out, although it did trigger a flurry of interest and some ongoing discussions via email (maybe even one video meeting but I can't remember the details now.)
Oct 8 2024
Any guess about a supposed right value for service? lol Just for my curiosity
Oct 2 2024
Just talking about accessibility, maybe interesting, Wikipedia on mobile allows to scale, but setting a minimum and maximum scale:
Thanks for the very detailed answer! I may or may not take a stab at this if the person who originally asked for it decides that this is worth the effort.
Oct 1 2024
Stumbled upon https://we.phorge.it/rP95e179d9a4f31b8c5c2cbb4db8f7fa9f2d3867d6 which also mentions
We may not have a commit object for a given identifier if the commit has not imported yet.
So this can indeed happen.
Sep 30 2024
Ah, hadn't even thought to check Developer Mode, I assumed I had it enabled. I did a "Purge Caches" from there and that didn't do anything, but turning devmode on did.
it might be in the APCu cache (https://we.phorge.it/config/cache/) - I'll try to check later. Also make sure you've set the Developer Mode in settings.
Sep 28 2024
Ok, so --branch probably doesn't make much sense in that case, right.
Sep 26 2024
In D25823#22041, @avivey wrote:mm, isBranchJustFilesystemPath() makes more sense now, as in "the branch name is part of the URI path"...
mm, isBranchJustFilesystemPath() makes more sense now, as in "the branch name is part of the URI path"...
Maybe I should just try to improve the ArcanistRepositoryAPI#getDefaultBranch() 🤔 since the root problem is there, and the current solution just involves more code to avoid to fallback on that "master" default
Sep 25 2024
done
move confusing comment
In D25823#22023, @avivey wrote:a. I'm not sure I like isBranchJustFilesystemPath() to check Is this SVN - if we're trying to be generic, then there may be a different vcs that does this particular trick, but has different logic; And if we're not trying to be generic, then isSVN() is probably a good-ish name? or isVCS('SVN').