Happened again even with max_wal_size=32GB on the 128GB WAL partition. At
this point I'm quite sure what happens:
+ a checkpoint is going on for several minutes (because of I/O contention
on the archival drive)
+ meanwhile data keeps coming in through several processes of pg_restore
+ data is
Hello list,
last night I got ENOSPACE on the WAL partition while running a huge
pg_restore on an empty and idle database.
The checkpoint that started 7 minutes earlier never finished:
04:31:09 LOG: checkpoint starting: wal
...
04:38:04 PANIC: could not write to file "pg_wal/xlogtemp.5462
Forgot an important setting I have during pg_restore:
On Thu, 10 Apr 2025, Dimitrios Apostolou wrote:
max_wal_size=64GB
max_replication_slots = 0
max_logical_replication_workers = 0
max_wal_senders = 0
wal_level = minimal
autovacuum = off
checkpoint_completion_target = 0
so the checkpoint sh
