English speaking PostgreSQL public chat. This group is for discussions on PostgreSQL-related topics and strives to provide best-effort support as well.
I really don't like the idea of having multiple OS users for each postgresql user.
Читать полностью…Please help me create the statefullset for postgres in kubernetes with replication n auto failover (HA)
Читать полностью…Hi Guys
Is there any script or something you guys know of. That script should scan my database, and see what are the columns that will be helpful to index but aren't indexed and show me them. I remember this was present as some script in sql server, but I want to know for postgresql is there is such
Wait, are we talking about "idle in transaction" or "idle" sessions?
Читать полностью…Those have the bad habit of immediately coming back. Talk to your developers about connection pool sizing instead.
Читать полностью…We are using Aurora postgres databases.. In the database i can see lot of idle transactions.. Is it okay? Should i delete?
Читать полностью…is very small hackathon(<$10k), but came 400+ participants, i have suspicion that organisers have problems of scale to determinate the winner (announced only 3 judges) 100+ works per judge is 2-3 weeks of work
Читать полностью…What is the best channel on YouTube to watch advance topics
Читать полностью…There might be other issues of course - anything that corrupts data in the OS or hardware could also corrupt an index.
Читать полностью…Perhaps you have upgraded the version of glibc(?) and that changed the sorting weights so that indexes are invalid? If so it is necessary to immediately reindex or else things like that can happen.
Читать полностью…(row or column-level filtering may require specific versions)
Читать полностью…You can of course make a physical replica and create a logical replica from it. Synchronisation of switching from that replica to the primary might be tricky but perhaps something could be done with slots if they exist that far back.
In my experience your best bet is to remove or exclude indexes from the structural backup and add any necessary after the initial replica is created. Importing large tables with many resulting index updates can be very slow vs. just creating the index after.
Hi I am looking for a course please let me know if any one
Читать полностью…What is the "proper" way to create new users and databases? I see conflicting tutorials online, some of them create a system (OS) user for a new postresql user, like here: https://wiki.debian.org/PostgreSql. others do not mention any os user and just create users and databases via postgresql's superuser
Читать полностью…AFAIK PostgreSQL doesn't track how often a certain column appears in a WHERE. A naïve idea would be to check your pg_stat_user_tables for significant SeqScan counts, then get the list of columns and query pg_stat_statements for queries havingquery ILIKE relname || ' WHERE %' || colname
orquery ILIKE relname || ' ON %' || colname
You get the idea...
I suppose that he refer to idle in transactions which are dangerous because they don't allow autovacuum to do its job
Читать полностью…It's not a good thing and something you should grt you app developers to fix. It implies your app is starting a transaction and then not ending it before returning it to some form of connection pool.
Читать полностью…Yes, I refuse. They do not even capitalise PostgreSQL consistently!
Читать полностью…Well, if you insert twice NOW (after the glibc switch) they will match each other.
Create a new table abc(LIKE orig_table including all);
Insert into abc select * from orig_table;
see if those inserts fail, if they do, it is LIKELY the glibc issue.
Hi everyone,
I participated in two hackathons on Devpost:
The Syrotech MVP Hackathon 2025
Eco-Innovation Challenge
The winner announcements were supposed to be on September 1st and 6th, respectively, but I still don’t see any winners posted.
Does anyone know if Devpost is legit or if this might be a scam?
Index is currently being used and is not invalid. If I insert the duplicate record it will not allow.
Читать полностью…Fixing it can be tricky, because you may need to target the invalid data without using an index. There are debug settings you can use to prevent e.g. index scans being used. https://www.postgresql.org/docs/15/runtime-config-query.html#RUNTIME-CONFIG-QUERY-ENABLE
Читать полностью…I found that postgres is having some problem with the primary key as it violate sometimes and duplicate records are created in the table. Has anyone else observed the same. I am using postgres 15.7. Is there any solution to this problem pls let me know.
Читать полностью…You could also exclude some data, possibly, if you don't need a complete replica for your purposes. Maybe some tables (or columns/rows within them) are unnecessary, merely containing transient data of temporary use.
Читать полностью…Question regard logical replication in version 10-13:
is it possible to set up the replication from pg_basebackup instead of waiting for the CREATE SUBSCRIPTION to start sync initial data?
what are my options to make the replica be ready as fast as possible?
Hi All ,
I am looking for Postgresql 17 pre and post installation standards on windows. Can any one share ref docs please..
You're welcome, albeit I don't remember giving input ;-)
Читать полностью…