*DEAR TOM*
just one PG instance in host I did an experiment When I remove pg and reinstall pg, the function of pg_hba is working ,represent that the location of pg_hba is right ----- remove yum remove postgresql* --- install yum -y install https://download.postgresql.org/pub/repos/yum/reporpms/EL-7-x86_64/pgdg-redhat-repo-latest.noarch.rpm yum install postgresql10 postgresql10-server postgresql10-contrib postgresql10-libs postgresql10-dev* -y ----------- I have Check again the content of pg_hba.conf and "select * from pg_hba_file_rules" consistent Yes, this question is very tricky Tom Lane <t...@sss.pgh.pa.us> 於 2021年12月21日 週二 下午10:42寫道： > shing dong <s7eqs...@gmail.com> writes: > > 1. The rules in pg_hba.conf are almost invalid > > 2. pg_hba.conf is only useful for METHOD = trust > > 3. check SHOW hba_file; the file location is correct > > 4. select * from pg_hba_file_rules; checked is correct > > 5.DB version : PostgreSQL 10.19 on x86_64-pc-linux-gnu, compiled by gcc > > (GCC) 4.8.5 20150623 (Red Hat 4.8.5-44), 64-bit > > Even if you delete the text in pg_hba.conf > > Keep only > > host VJ VJ_USER 10.10.10.1/32 md5 > > After pg_ctl reload and Restart DB , any ip, user still can log in to > DB > > It's hard to say where your mistake is, but probably the first > thing to check is whether you're really restarting the postmaster. > I'm wondering in particular if there's more than one PG instance > on the machine and you're reconfiguring or restarting the wrong > one. Other than that, retrace your steps carefully, because at > least one of the above statements must be wrong. > > (I guess if you were feeling *really* paranoid, you could wonder > whether somebody replaced your postmaster executable with a hacked > version that doesn't apply any pg_hba checks. But pilot error > seems like a far more probable explanation.) > > regards, tom lane >
