| hoo added a comment. |
Given the lock timeouts happen on SELECT, I don't really see how using a single query instead of SELECT+UPDATE/INSERT would greatly help us here. While that's definitely nice to have performance wise (less queries, less roundtrips), the main problem seems to me that something is getting a new id and then keeping the transaction (=> the lock) open to do other things before commit/rollback.
While I definitely think that using UpsertSqlIdGenerator is an improvement and will greatly help us here once the contention around this table grows even more… I don't think it can fix the problem at hand. We wont be able to solve the problem as long as we don't commit immediately after UPDATE/INSERT, I fear.
TASK DETAIL
EMAIL PREFERENCES
To: Addshore, hoo
Cc: Ladsgroup, aaron, TerraCodes, Stashbot, Banyek, Paladox, mmodell, daniel, Addshore, Aklapper, gerritbot, hoo, CucyNoiD, Nandana, NebulousIris, Gaboe420, Versusxo, Majesticalreaper22, Giuliamocci, Adrian1985, Cpaulf30, Lahi, Gq86, Baloch007, Darkminds3113, Bsandipan, Lordiis, GoranSMilovanovic, lisong, Adik2382, Th3d3v1ls, Ramalepe, Liugev6, QZanden, LawExplorer, Lewizho99, Maathavan, _jensen, D3r1ck01, Jonas, Wikidata-bugs, aude, Lydia_Pintscher, Jdforrester-WMF, Mbch331, Jay8g, Krenair
Cc: Ladsgroup, aaron, TerraCodes, Stashbot, Banyek, Paladox, mmodell, daniel, Addshore, Aklapper, gerritbot, hoo, CucyNoiD, Nandana, NebulousIris, Gaboe420, Versusxo, Majesticalreaper22, Giuliamocci, Adrian1985, Cpaulf30, Lahi, Gq86, Baloch007, Darkminds3113, Bsandipan, Lordiis, GoranSMilovanovic, lisong, Adik2382, Th3d3v1ls, Ramalepe, Liugev6, QZanden, LawExplorer, Lewizho99, Maathavan, _jensen, D3r1ck01, Jonas, Wikidata-bugs, aude, Lydia_Pintscher, Jdforrester-WMF, Mbch331, Jay8g, Krenair
_______________________________________________ Wikidata-bugs mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikidata-bugs
