Обсуждение: BUG #15292: Multiple Messages "2018-07-24 10:46:12.566 CEST [32029]LOG: invalid length of startup" in logfile
BUG #15292: Multiple Messages "2018-07-24 10:46:12.566 CEST [32029]LOG: invalid length of startup" in logfile
От
PG Bug reporting form
Дата:
The following bug has been logged on the website: Bug reference: 15292 Logged by: stefan muster Email address: stefan.muster@isc-ejpd.admin.ch PostgreSQL version: 11beta2 Operating system: SUSE Linux Enterprise Server 12 (x86_64) Description: Modification in pg_hba.conf: Add line "host all all 172.24.84.43/0 md5" in Section # IPv4 local connections: And Modification in postgresql.conf: listen_addresses = '*' After Restart postgres Cluster, I receive all Minutes Messages like ""2018-07-24 10:46:12.566 CEST [32029] LOG: invalid length of startup" in Logfile. But the remote Connection from Host 172.24.84.43 is working as expected. If listen_addresses = '*' change to listen_addresses = 'localhost' and restart the Postgres Cluster, I do not receive any further error messages in logfile. But as expected, the remote connections are not working in this case. I've also postgresql version 10 on the same host and can not detect any issuses when we change listen_adresses = '*' Thanks a lot for your Investigation and feedback. With best regards Stefan
=?utf-8?q?PG_Bug_reporting_form?= <noreply@postgresql.org> writes: > Modification in pg_hba.conf: > Add line "host all all 172.24.84.43/0 > md5" in Section # IPv4 local connections: The above is accepting connections from *anywhere*, not only 172.24.84.43. ITYM 172.24.84.43/32 ... > And Modification in postgresql.conf: > listen_addresses = '*' > After Restart postgres Cluster, I receive all Minutes Messages like > ""2018-07-24 10:46:12.566 CEST [32029] LOG: invalid length of startup" in > Logfile. But the remote Connection from Host 172.24.84.43 is working as > expected. What that says is you're getting port-scanned by something or other, once you expose a PG port on your network-accessible interface. regards, tom lane