> On Mar 26, 2019, at 8:08 AM, Adrian Klaver <adrian(dot)klaver(at)aklaver(dot)com> wrote:
>
> On 3/25/19 5:10 PM, Rob Sargent wrote:
>> PG10.7, Centos7
>> On Mar15 we filled our default tablespace/WAL partition. Cleaned up some old dumps and restarted. pg_wal had apparently exploded but cleaned itself up by the next day. On Mar16 I ran CHECKPOINT in all databases on that server (except template0). All seems fine except for 271 WALs from MAR16 going nowhere fast. Of course we see transient WALS come and go every day. We did a restart of the server last Friday (Mar22) without issue.
>> -rw-------. 1 postgres postgres 16777216 Mar 16 16:32 0000000100000CE9000000DD
>> ...(269 others)
>> -rw-------. 1 postgres postgres 16777216 Mar 16 17:01 0000000100000CEA000000E9
>> What’s keeping these alive. Or can they be deleted? Any normal admin routine likely to clean these up?
>
> wal_keep_segments?
Commented out in postgres.conf
>
> Do you have replication set up and replication slots in use?
No, sorry I should have said that up front. We’re simple folk.