Re: Cygwin cleanup - Mailing list pgsql-hackers

From Justin Pryzby
Subject Re: Cygwin cleanup
Date
Msg-id 20230126211511.GU22427@telsasoft.com
Whole thread Raw
In response to Re: Cygwin cleanup  (Justin Pryzby <pryzby@telsasoft.com>)
List pgsql-hackers
Note that cirrus failed like this:


https://api.cirrus-ci.com/v1/artifact/task/4881596411543552/testrun/build/testrun/subscription/010_truncate/log/010_truncate_publisher.log

2023-01-25 23:17:10.417 GMT [29821][walsender] [sub1][3/0:0] ERROR:  could not open file
"pg_logical/snapshots/0-14F2060.snap":Is a directory
 
2023-01-25 23:17:10.417 GMT [29821][walsender] [sub1][3/0:0] STATEMENT:  START_REPLICATION SLOT "sub1" LOGICAL 0/0
(proto_version'4', origin 'any', publication_names '"pub1"')
 
2023-01-25 23:17:10.418 GMT [29850][walsender] [pg_16413_sync_16394_7192732880582452157][6/0:0] PANIC:  could not open
file"pg_logical/snapshots/0-14F2060.snap": No such file or directory
 
2023-01-25 23:17:10.418 GMT [29850][walsender] [pg_16413_sync_16394_7192732880582452157][6/0:0] STATEMENT:
START_REPLICATIONSLOT "pg_16413_sync_16394_7192732880582452157" LOGICAL 0/14F2060 (proto_version '4', origin 'any',
publication_names'"pub3"')
 

I don't understand how "Is a directory" happened ..

It looks like maybe the call stack would've been:

SnapBuildSerializationPoint()
xlog_decode() or standby_decode() ?
LogicalDecodingProcessRecord()
XLogSendLogical()
WalSndLoop()
StartLogicalReplication()

-- 
Justin



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: suppressing useless wakeups in logical/worker.c
Next
From: Robert Haas
Date:
Subject: Re: New strategies for freezing, advancing relfrozenxid early