I manage two Postgres instances (12.14 and 14.7) on a Windows 10 server at work. Both instances are regular EnterpriseDB installations. Lately, I see the following log messages on both instances: 2023-03-07 02:37:08.171 CET [4624] LOG: could not rename temporary statistics file "pg_stat_tmp/global.tmp" to "pg_stat_tmp/global.stat": Permission denied 2023-03-07 02:37:11.874 CET [3656] LOG: using stale statistics instead of current ones because stats collector is not responding Both messages appear at irregular intervals. The first message every few minutes. The second message every few hours. The "permission denied" message apparently is caused by another process having an open file handle at the same time, likely some antivirus software scanning the system. What's the impact of this on Postgres, besides using stale statistics, if this does not get resolved? I assume the file will be renamed eventually once no other process has an open file handle and the statistics will eventually be refreshed. The log messages appeared after a server admin switched from Symantec Endpoint Protection to Cortex XDR. I had the admin exclude Postgres from the scan and they disabled Cortex XDR entirely. But the messages still appear, although less frequently. I doubt that Cortex XDR is fully disabled because it's still writing its event log. But its UI says disabled. The switch to Cortex XDR was triggered because something deleted pg_ctl.exe from my Postgres 14 instance. I still don't know what caused that. But the admin had the brilliant idea of replacing one snake oil with another. Maybe it's unrelated to the snake oil but I haven't received any further info on possible causes from the admin. -- Erik