I have just built the deploy repo (cf above patch). Have you rebuilt your node module deps before building the deploy repo?
Sat, Jun 15
Fri, Jun 14
Thu, Jun 13
As of T180384: Turn off Trending Service the trending edits service is no longer a production service.
Wed, Jun 12
Great to hear!
We have recently introduced storage groups for the SQLite back-end but haven't updated the example config, sorry about that! What you need to do is to add the following under the table configuration:
Tue, Jun 11
Mon, Jun 10
This is a huge bot-edited gallery page. For performance reasons we don't support such requests (as it takes more than 120 seconds for Parsoid to render it).
Sat, Jun 8
This is the fault of the error-handling code in the service template. PR #122 should fix this.
Yup, this is due to the missing ca-cetrificates package. It really depends on a per-service basis: if and only if the service is meant to communicate with external sources, then the package should be installed. In the case of the service template, I think we should install it only for the development and test images, but not the production one in an effort to lock down services using the blubber file as a template.
All of the formulae have been rechecked, so warnings should now be emitted for all requests. Since we updated Mathoid to the latest code, I have also regenerated all of the renders to match the current expectations.
Fri, Jun 7
Yes, we could, but likely with an ETA of 2 months. When you say MCS, does that include all of MCS-backed end points?
Thu, Jun 6
Officewiki si a private wiki, and we don't have RESTBase set up for private wikis.
Image building is now working, so resolving. I created T225236: Figure out how to test Citoid with Zotero in the pipeline where we can discuss how to properly test Citoid with Zotero inside CI.
The first part of the issue has been fixed; Mathoid's code is up to date now. However, the results will only be publicly visible for new formulae. I'll run a dump of all of the formulae in storage next week.
I think the root cause is the same as T208922: PHP Fatal Error: Class undefined: JobExecutor (jobrunners try to run labswiki jobs), namely T208922#4766050 (global config problems when spawning inter-wiki jobs). It's just that now we detect the case when JobExecutor is not loaded instead of bluntly crashing.
Thank you, @thcipriani for the investigation. In the above patch I updated the deps that were known to cause issues with 504s, but the build still fails. Could it be that Citoid can't talk to Zotero any more?
Wed, Jun 5
I am not sure I understand your problem. When you open the Math docs section, there is a Try it out button. Once you press it and enter the formula to check, click on Execute and the full curl command is displayed, together with the request URL and the response returned by the REST API.
There's something weird in the Mathoid deployment here. When I issue the same request from inside production directly to Mathoid, the warnings are not returned either:
Tue, Jun 4
This is an interesting idea. The issue here, though, is: how can updates to such metadata be triggered?
The RESTBase switch has been deployed. Proton is now used for all PDF requests.
Sat, Jun 1
The new proposed end points LGTM.
Fri, May 31
Thu, May 30
We have settled on restrouter and restbase for simplicity.
Both Parsoid and MCS storage are now using only binary key/value buckets for their storage needs and key_rev_value and multi_bucket are not used anymore, so this task is now complete.
Everything here has now been completed \o/ Resolving.
The corresponding keyspaces have been dropped. Resolving.
We have switched to OpenAPI v3 (together with the accompanying swagger-ui library update), so this should not be an issue any more. Please reopen if that's not the case.
We have switched to the new swagger-ui version and OpenAPI v3. If you perceive the problem again, please reopen.
This has now been deployed.
Wed, May 29
For restbase-dev1006 there is T224260: restbase-dev1006 has a broken disk, so that would need to be resolved first (and AFAIK it's likely we'll need to end up re-imaging the machine anyway).
Agreed. We can start coordinating once T208087: Replace remaining Samsung SSDs is done. That includes T223976: Decommission restbase10(0[7-9]|1[0-5]) but also T222960: Fix restbase1017's physical rack.
Moving back to backlog due to inactivity.
Moving to backlog until @Jdlrobson is back.
Last step: deployment
Tue, May 28
We can also just add the header, since it is ignored by Parsoid/JS any way