joined node can have extra databases

Bug #540187 reported by Seppo Jaakola
10
This bug affects 2 people
Affects Status Importance Assigned to Milestone
MySQL patches by Codership
New
Low
Seppo Jaakola
0.7
Won't Fix
Low
Seppo Jaakola
0.8
Won't Fix
Low
Seppo Jaakola

Bug Description

If a joining node has some MySQL databases, which are not present in the cluster, then after The DB SST, the joiner will have all cluster databases + his private databases. This happens because the SST procedure does not clean the joiner's databases, it just loads new dump from donor.
This can cause an issue, if some client trx, connecting to the joiner node, references tables in the private database. Appliers in other cluster nodes will fail for references into non-existing database.

Changed in codership-mysql:
importance: Undecided → Medium
importance: Medium → Low
assignee: nobody → Seppo Jaakola (seppo-jaakola)
milestone: none → 0.7.4
Changed in codership-mysql:
milestone: 0.7.4 → none
summary: - joiner can have excessive databases
+ joined node can have extra databases
Revision history for this message
shinguz (oli-sennhauser) wrote :

Donor terminates itself when joiner is inserting rows into local databases on joiner which are not existing on the donor.

No good for operations! Joiner should be cleaned completely by default. Either by mysql_install_db or by "DROP ALL DATABASES";

Release: 23.2.4(r147)

130410 15:28:06 [Warning] WSREP: BF applier failed to open_and_lock_tables: 1146, fatal: 0 wsrep = (exec_mode: 1 conflict_state: 0 seqno: 22516)
130410 15:28:06 [ERROR] Slave SQL: Error executing row event: 'Table 'trullalla.customer' doesn't exist', Error_code: 1146
130410 15:28:06 [Warning] WSREP: RBR event 2 Write_rows apply warning: 1146, 22516
130410 15:28:06 [ERROR] WSREP: Failed to apply trx: source: 25291aff-a1e1-11e2-0800-40fa58d6fa54 version: 2 local: 0 state: APPLYING flags: 1 conn_id: 10 trx_id: 119596 seqnos (l: 22580, g: 22516, s: 13262, d: 21523, ts: 1365600381059020151)
130410 15:28:06 [ERROR] WSREP: Failed to apply app buffer: seqno: 22516, status: WSREP_FATAL
         at galera/src/replicator_smm.cpp:apply_wscoll():53
         at galera/src/replicator_smm.cpp:apply_trx_ws():120
130410 15:28:06 [ERROR] WSREP: Node consistency compromized, aborting...

Greetings from Galera Cluster Training at LinuxHotel by FromDual! :-)
Sebastian, André, Phillip, Dennis and Oli

Revision history for this message
Alex Yurchenko (ayurchen) wrote :

Oli, have you been using mysqldump SST?

Revision history for this message
Raghavendra D Prabhu (raghavendra-prabhu) wrote :

This is similar to https://bugs.launchpad.net/percona-xtradb-cluster/+bug/1213073 but that was for Xtrabackup SST. This shouldn't affect rsync SST currently, so is this bug for mysqldump SST (which is no longer the default)?

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.