3 months ago
Hey everyone, getting this issue on all routes:
psycopg2.errors.DiskFull: could not resize shared memory segment "/PostgreSQL.2188626702" to 8388608 bytes: No space left on device
How can I reset the memory or something, our prod db is failing right now on a lot requests, its a pgvector instance, i tried redeploying and restarting
7 Replies
I fixed it by setting a variable:
RAILWAYSHMSIZE_BYTES="512000000"
Is this the correct fix? Want to figure out if this is application level issue or railway
If it's the correct fix, as I've found other threads saying so, does it increasingly occurr? As in could there be a memory leak somewhere that even with the increased bytes it will crash in like a few weeks
Also unrelated, is it possible to migrate the database to another project?
Also are rollbacks no immediate? Our API crashed and I rolled back to previous and it's been building for 6 minutes