Stand by server (9.6.6) with corrupt file

Environment: PostgreSQL 9.6.6 installed from yum repository. Oracle Linux 7 EL x64. Dell servers with Raid 5 (hw).

I was testing our database backup system (based on pgBarman), and discovered that one base file is corrupt in our standby database server. The file is OK in master server, but has 0 bytes in size in standby server.

Looking master and standby servers there is no indication that the problem exists - replication is running fine.

Evidences:

On master server:

On standby server:

After long investigation, I discovered that if I execute a query on standby server:

< 2017-12-22 11:20:22.417 -02 > ERROR:  could not read block 0 in file "base/1106839/6302536": read only 0 of 8192 bytes
< 2017-12-22 11:20:22.417 -02 > STATEMENT:  SELECT *
          FROM MY_FAIR_LARGE_TABLE t1

         LEFT OUTER JOIN MY_FAIR_LARGE_SUBTABLE t0 ON (t0.the_id = t1.ID)
                   WHERE (((t1.COMPANY_ID = 2)
                     AND t1.OTHERCOMPANY LIKE '20147617%')
                     AND (t1.TEST_FLAG = 0))
        ORDER BY t1.DUE_DATE LIMIT 1000 OFFSET 0

Very same query on server works fine.

And there is no replication error - everything is in sync between these two servers (I know, I'm begin to be repetitive).

I've about 30 servers with same setup, and this only has this flaw. The only difference is that this database is about 3 times larger than the others (about 90Gb in size).

Server and slave have 23ms of network lag - which seems not be a problem for the other databases in the same server.

Any advice?

--

Edson Carlos Ericksson Richter SimKorp Ltda
Fone:	(51) 3366-7964
"A mente que se abre a uma nova ideia jamais voltará ao seu tamanho original" - Albert Einstein