Previously, if the witness node's PostgreSQL was unreachable, it would
be marked as "inactive" on the primary, and under some circumstances
would not be corrected to "active" once the witness node's PostgreSQL
came back.
PR #754; some modifications by Ian Barwick.
The cluster size check is purely informative, and is not in any way
essential for the standby clone operation. As it's possible the query
may fail if the repmgr user does not have sufficient privileges to
query all databases in the cluster, we can simply ignore any failure.
Note that the code comment indicated the query also served to sanity-
check that queries can actually be executed. While this was the case
historically, the preceding server version check now serves the same
purpose and will not have the same risk of failure due to missing
permissions.
A number of C functions were added in releases 4.2 to 4.4; however
these were renamed in 5.3 to prevent naming clashes with other
extensions.
This does however mean that when upgrading from one of the above
versions, the intermediate upgrade steps will attempt to create
SQL functions referencing C functions which no longer exist in
repmgr.so, and hence cause the upgrade to fail.
We can work around this by providing empty upgrade scripts
from these versions to 4.4, which skip the problematic CREATE
FUNCTION commands. The functions will be correctly created in
the 5.2--5.3 upgrade script.
When recovering from degraded state in local node monitoring, in some
cases a new connection was opened to the local node without closing
the old one, which will result in memory leakage.
Remove bogus -W option in "repmgr standby follow" example invocation
for the "follow_command" parameter.
The option (which corresponds to "--no-wait") is not used by
"repmgr standby follow".
Per report from Jimmy Angelakos.
Commit 79d1f00 modified repmgrd to automatically set an inactive node
to "active" at startup.
However we need to avoid doing that for cases where the node role has
changed (e.g. a former primary was recloned as a standby) but the node
record was not updated.
From PostgreSQL 9.4 (commit e7128e8d), explicit function prototypes
are not required as they will be generated by the PG_FUNCTION_INFO_V1
macro.
(We were supporting PostgreSQL 9.3 until relatively recently, so there
was nothing particular to gain by removing these earlier).
Explicitly check whether the "repmgr node rejoin" command on the
demotion candidate succeeded. Due to the way SSH execution is
currently implemented, we can return either the command execution
status or the command output; to ensure any errors are available,
log them to a temporary file on the demotion candidate and note
its location in case of an error.
While we're at it, improve error message handling when the demotion
candidate fails to rejoin.
