Details
-
Bug
-
Status: On Hold
-
Medium
-
Resolution: Unresolved
-
8.0.20-11.2(CVE)
-
None
-
None
-
None
-
To the environment and versions on all server exactly the same:
CentOS Linux release 8.2.2004 (Core)
xtrabackup version 8.0.14 based on MySQL server 8.0.21 Linux (x86_64) (revision id: 113f3d7)
/usr/sbin/mysqld Ver 8.0.19-10 for Linux on x86_64 (Percona XtraDB Cluster (GPL), Release rel10, Revision 727f180, WSREP version 26.4.3)
To the environment and versions on all server exactly the same: CentOS Linux release 8.2.2004 (Core) xtrabackup version 8.0.14 based on MySQL server 8.0.21 Linux (x86_64) (revision id: 113f3d7) /usr/sbin/mysqld Ver 8.0.19-10 for Linux on x86_64 (Percona XtraDB Cluster (GPL), Release rel10, Revision 727f180, WSREP version 26.4.3)
Description
We reduced our 5 Node 5.7 Cluster to only one node. Then we upgraded this node to PXC 8 by the following steps:
- stopped the node and upgraded all the packages
- applied all the configchanges for 8
- made the node readonly and started it without the galera configurations
- node performed an inplace upgrade, everything was fine
- executed a "reset master" on the node and stopped it again
- removed the readonly from config and added the wsrep configs to the node again (with empty gcomm address)
- started the node again and we had our frist PXC 8 node with galera
Now we was going on to the next node:
- mysqld was already stopped
- upgraded the packages to pxc8 (including xtrabackup as on the first node)
- we applied the configchanged for 8
- then we cleared the datadir by "rm -rf DATADIR/*"
- then we started the node
- the node recognized the empty directory and started an SST from our first node
- this took 3h since its 1,4TB
- At the end the receiver node throwed the following errors:
Sep 02 23:50:31 receiver-node.example.com -wsrep-sst-joiner[24784]: 2020-09-02T21:50:25.614581Z 1 [System] [MY-000000] [WSREP] PXC upgrade completed successfully
Sep 02 23:50:31 receiver-node.example.com -wsrep-sst-joiner[24784]: mysqld: File '/data/mysql/.000003' not found (OS errno 2 - No such file or directory)
Sep 02 23:50:31 receiver-node.example.com -wsrep-sst-joiner[24784]: 2020-09-02T21:50:25.680151Z 0 [ERROR] [MY-010958] [Server] Could not open log file.
Sep 02 23:50:31 receiver-node.example.com -wsrep-sst-joiner[24784]: 2020-09-02T21:50:25.680445Z 0 [ERROR] [MY-010041] [Server] Can't init tc log
Sep 02 23:50:31 receiver-node.example.com -wsrep-sst-joiner[24784]: 2020-09-02T21:50:25.681044Z 0 [ERROR] [MY-010119] [Server] Aborting
Sep 02 23:50:31 receiver-node.example.com -wsrep-sst-joiner[24784]: 2020-09-02T21:50:30.620730Z 0 [System] [MY-010910] [Server] /usr/sbin/mysqld: Shutdown complete (mysqld 8.0.>
Sep 02 23:50:31 receiver-node.example.com -wsrep-sst-joiner[24784]: EOF:
2020-09-02T21:50:32.116964Z 0 [ERROR] [MY-000000] [WSREP] Process completed with error: wsrep_sst_xtrabackup-v2 --role 'joiner' --address '192.168.0.201' --datadir '/data/mysql/' --basedir '/usr/' --plugindir '/usr/lib64/mysql/plugin/' --defaults-file '/etc/my.cnf' --defaults-group-suffix '' --parent '24259' --mysqld-version '8.0.19-10' --binlog '/data/mysql/mysql-bin' : 3 (No such process)
2020-09-02T21:50:32.119162Z 0 [ERROR] [MY-000000] [WSREP] Failed to read uuid:seqno from joiner script.
2020-09-02T21:50:32.119207Z 0 [ERROR] [MY-000000] [WSREP] SST script aborted with error 3 (No such process)
Sep 02 23:50:33 receiver-node.example.com systemd[1]: mysql.service: Main process exited, code=killed, status=6/ABRT
Sep 02 23:50:33 receiver-node.example.com systemd[1]: mysql.service: Failed with result 'signal'.
Sep 02 23:50:33 receiver-node.example.com systemd[1]: Failed to start Percona XtraDB Cluster.
for any reason, the node replicates the binlog file with the correct name to the data directory (/data/mysql/mysql-bin.000003) However in the index file (/data/mysql/mysql-bin.index) the wrong filename is written (/data/mysql/.000003).
So why is there a completly wrong filename in the index and how could I solve this issue? I tried to simply start the node again and hoped it will recognize this error and start an IST, however it just dropped everything and started a new SST.
Attachments
Issue Links
- relates to
-
PXB-2451 PXB 8.0 backup fails with "Error cannot open binlog index file" if periods in filename
-
- Done
-