Issues

• Active-passive MariaDB/Percona/MySQL is pinned to one node on a local volume: no failover volume

  active-passivebugopensvc
  signal18/replication-manager#2004 · Filter by author svaroqui opened on Oct 12, 2026

• ColumnStore child cluster: 1 UM + x PM, each PM an active-passive DRBD service, replicating from an OLTP parent

  configuratorenhancementopensvc
  signal18/replication-manager#2002 · Filter by author svaroqui opened on Oct 12, 2026

• server test TestHandlerMuxServersPortConfig_ValidBasicAuthPassesACL panics: nil Conf in apiTokensEnabled

  apibug
  signal18/replication-manager#2000 · Filter by author svaroqui opened on Oct 12, 2026

• Logical backup of a MySQL/Percona GTID replica writes errant transactions on it (--dump-slave)

  backupbuggtidmysql
  signal18/replication-manager#1996 · Filter by author caffeinated92 opened on Oct 12, 2026 · v3.1.44

• ssl DB tag breaks MySQL/Percona 8.4: the rendered bare 'ssl' server option was removed in 8.4

  bugconfiguratormysqltls
  signal18/replication-manager#1993 · Filter by author caffeinated92 opened on Oct 12, 2026 · v3.1.44

• OpenSVC v3 orchestrated start treats 409 'already started' as a failure: the rolling reprov cancels at its first node

  bug
  signal18/replication-manager#1981 · Filter by author svaroqui opened on Oct 11, 2026

• splitdump.ClassifyStream puts the dump footer (SET ...=@OLD_*) into the system artifact without its header: system-catalogue replay fails 1231 (time_zone NULL)

  bug
  signal18/replication-manager#1980 · Filter by author svaroqui opened on Oct 11, 2026

• Reseed restores the backup's raw mysql system tables: breaks a node on another release; restore strategy must depend on a fresh-node detection

  bug
  signal18/replication-manager#1979 · Filter by author svaroqui opened on Oct 11, 2026
  2

• GetConnNoBinlog returns its sql_log_bin=0 session to the server.Conn pool: later writes on the master are not binlogged, silently out of replication

  bug
  signal18/replication-manager#1978 · Filter by author svaroqui opened on Oct 11, 2026

• dev3 (MariaDB 12.3.3): GTID binlog state emptied on both masters around switchovers; after a rolling-restart switchback the replicas fail 1236

  bug
  signal18/replication-manager#1977 · Filter by author svaroqui opened on Oct 11, 2026

• Dropping a DB tag does not revert it: the delta keeps the tag's variables as delta:no-config in 02_delta.cnf

  bug
  signal18/replication-manager#1976 · Filter by author svaroqui opened on Oct 11, 2026

• DB tag matched as a suffix: the ssl tag also applies strictssl (require_secure_transport=ON live, monitoring locked out); bind matches nobind, statement matches nologslavestatement

  bug
  signal18/replication-manager#1975 · Filter by author svaroqui opened on Oct 11, 2026

• Database DSN not rebuilt when the TLS certificates change: dropping the ssl tag leaves repman unable to connect (invalid connection)

  bug
  signal18/replication-manager#1974 · Filter by author svaroqui opened on Oct 11, 2026

• Cluster add: admin left a visitor (no Switchover/Failover) and the system account dropped until the next restart

  apibugregressionsecurity
  signal18/replication-manager#1966 · Filter by author svaroqui opened on Oct 10, 2026

• Thread pool: lower thread_pool_stall_limit when the pool gets close to its limit (new connections wait 500 ms at 60% full)

  signal18/replication-manager#1961 · Filter by author svaroqui opened on Oct 9, 2026

• Degraded-link test profile: replication-manager over satellite-class links (latency, jitter, loss, handover cuts)

  signal18/replication-manager#1957 · Filter by author svaroqui opened on Oct 9, 2026

• /api/heartbeat returns arbitration-external-secret to any caller

  signal18/replication-manager#1955 · Filter by author svaroqui opened on Oct 9, 2026

• Never size the redo log above the page cache left to it (fscache share)

  signal18/replication-manager#1951 · Filter by author svaroqui opened on Oct 9, 2026

• PostgreSQL logical promotion does not synchronize sequences, causing serial/identity inserts to fail on the new publisher

  PostgreSQLreplication
  Bug

  signal18/replication-manager#1922 · Filter by author caffeinated92 opened on Oct 8, 2026

• PostgreSQL logical switchover leaves ddl_log sequence behind, causing DDL failures on the promoted publisher

  master slavePostgreSQLreplication
  Bug

  signal18/replication-manager#1921 · Filter by author caffeinated92 opened on Oct 8, 2026
  1

• Failover app on a DRBD volume: S3-mount init containers fail to provision on standby nodes and the dead peers' IP leases poison the cluster DNS (503 on the route)

  bugopensvc
  signal18/replication-manager#1918 · Filter by author svaroqui opened on Oct 7, 2026

• App route update does not run the gateway merge task: new routes stay 503 until task#mergecfg is run by hand

  apibughaproxy
  signal18/replication-manager#1917 · Filter by author svaroqui opened on Oct 7, 2026

• App sensor sidecar logs in as system at every push and each login triggers a failing SSO upgrade: 2 600 warnings per hour

  apibuglogs
  signal18/replication-manager#1912 · Filter by author svaroqui opened on Oct 7, 2026

• Standby automatic failover on a DNS blip, then the active forgets the master on an empty slaves list (om3 daemon restart, 2026-10-07)

  arbitratorbugfailoverrejointopology
  signal18/replication-manager#1909 · Filter by author svaroqui opened on Oct 7, 2026

• Gateway bandwidth fair share: when the shared uplink saturates, throttle the top cluster's app containers (cgroup / tc) to equalize the bandwidth per cluster

  enhancement
  signal18/replication-manager#1880 · Filter by author svaroqui opened on Oct 4, 2026