User Details
- User Since
- Mar 29 2015, 4:07 PM (468 w, 1 d)
- Availability
- Available
- LDAP User
- Tarrow
- MediaWiki User
- T Arrow (WMDE) [ Global Accounts ]
Thu, Mar 14
Wed, Mar 13
This is probably something worth looking at the next time we want to use a sitenotice for something
Tue, Mar 12
notes from a call about this:
Mon, Mar 11
We agreed last week to go ahead and close this since it now seems rather too late to make doing and CSS changes etc. valuable
Thu, Mar 7
Great, in that case I'm going to close this ticket. If you think I did this in error please feel free to reopen
I strongly suspect that this could be tied to the issues with search: how I would expect the selected items to be found. I tried this myself as another data-point and could replicate the behaviour shown in @Fring 's video (i.e. works for me)
We suspect this may have been due to our elasticsearch outages? @Charlie_WMDE / @So9q would you mind trying again?
Mon, Mar 4
Wed, Feb 28
Mon, Feb 26
This doesn't require some complete review or approval. I just wanted to stick this somewhere for another person to glance over before it's moved to done and resolved
Fri, Feb 23
This can now be resolved; we've now totally rebuilt all QS long since this happened and we have tooling for complete rebuilds now
Have you ever noticed a similar issue on wikidata.org or if this is just on wikibase.cloud?
Hi @Fnielsen, could to look and see if this is still an issue for you? I've had a little look and it appears to be resolved; feel free to reopen if some issues on your wikibase remain.
Hi @Skybristol could you take a look and see if these issues appear to be fixed for you?
Could you take a little look and see how the queryservice is looking for you now? Over the last months we've made a number of improvements to the reliability of the queryservice and also rebuild all wikibases queryservices.
Clearly this is a long time after you reported the issues but after we've put extensive work into improving the reliability of the queryservice and rebuilding the index I was wondering if you (and maybe @Drjwbaker) could comment if these issues now appear to be fixed? I'm actually going to mark the ticket as resolved but please don't hesitate to reopen if you believe that they are still present.
and this is now done; final job reun after using the following cleanup script: https://github.com/wmde/wbaas-deploy/blob/main/bin/queryservice-delete-dump-timestamp-in-namespaces
Thu, Feb 22
@Charlie_WMDE we heard form the grapevine that we're fine for this data to all be exposed publicly; is that right? Our technical implementation of it somewhat depends on this. For example, if we add this to the conversionMetrics "anyone" could then see the reason a user deleted a site; maybe that's desired but we wanted to check you were aware that is the case.
Wed, Feb 21
What this needs adding is a nice wikitext snippet for the actual banner content
Tue, Feb 20
down to just one wiki failing. https://ld4goats.wikibase.cloud/ with
java.lang.IllegalStateException: RDF store reports the last update time is before the minimum safe poll time. You will have to reload from scratch or you might have missing data. at org.wikidata.query.rdf.tool.change.ChangeSourceContext.getStartTime(ChangeSourceContext.java:100) at org.wikidata.query.rdf.tool.WbStackUpdate.initialize(WbStackUpdate.java:365) at org.wikidata.query.rdf.tool.WbStackUpdate.runUpdaterWithArgs(WbStackUpdate.java:287) at org.wikidata.query.rdf.tool.WbStackUpdate.main(WbStackUpdate.java:140) Failed batch: RDF store reports the last update time is before the minimum safe poll time. You will have to reload from scratch or you might have missing data. java.lang.IllegalStateException: RDF store reports the last update time is before the minimum safe poll time. You will have to reload from scratch or you might have missing data. at org.wikidata.query.rdf.tool.change.ChangeSourceContext.getStartTime(ChangeSourceContext.java:100) at org.wikidata.query.rdf.tool.WbStackUpdate.initialize(WbStackUpdate.java:365) at org.wikidata.query.rdf.tool.WbStackUpdate.runUpdaterWithArgs(WbStackUpdate.java:287) at org.wikidata.query.rdf.tool.WbStackUpdate.main(WbStackUpdate.java:140) Failed to run update command.
Mon, Feb 19
I basically exactly repeated this process (after cleaning up any errant jobs) but set -Xmx4g which should allow the heap of each of these jobs to rise to 4gb
Feb 16 2024
Doubling the memory got 86 done but 34 still fail due to memory.
Feb 15 2024
I've deleted all the successfully finished jobs:
kubectl -n qs-jobs delete jobs --field-selector status.successful=1
Feb 14 2024
Could be useful to have information on how often these links are clicked; if it's very few then we might want to prioritise this.
I checked this on wikibase cloud and it now errors when you try to enter a domain with a space
@Tarrow to check if this is still actually an issue; I think this is now fixed since we sorted out the unicode chars in domains
Feb 13 2024
Feb 12 2024
We thought we should ask you (@Charlie_WMDE) but maybe it's also meaningful to split by the usernames who did the deleting. For example more than 95% of the Jan 2023 deletion spike was due to a single user
Looks like we have 120 jobs that have stalled or failed:
$ kubectl -n qs-jobs get jobs -o go-template='{{printf "%d\n" (len .items)}}' --field-selector status.successful=0 120
Feb 9 2024
Feb 8 2024
From my perspective updating to PHP8.2 would make sense; I'm slightly unclear how large the work for doing this would be but I suspect not too great.
Feb 7 2024
When @conny-kawohl_WMDE and @Charlie_WMDE looked at this we decided that this ought to be timeboxed; we've moved it back to let engineers decide on a sensible timebox for this work in a refinement session.
Given that we haven't needed this for quite a while we are de-prioritising this. However if we have another extended Elasticsearch outage we should reconsider and maybe make this higher priority again.