Skip to:

  1. Skip to Jira Navigation
  2. Skip to Side Navigation
  3. Skip to Main Content

PXC-8.0 xtrabackup SST is failing with OS errno 32 - Broken pipe

Description

https://perconadev.atlassian.net/browse/PXC-8#icft=PXC-8.0 xtrabackup SST is failing with OS errno 32 - Broken pipe. The node will be automatically up and running after a few such failures.

Testcase

1) Initialize PXC-5.7 data directories ./pxc57/bin/mysqld --no-defaults --initialize-insecure --datadir=/dev/shm/qa/node1 --basedir=/dev/shm/qa/pxc57 ./pxc57/bin/mysqld --no-defaults --initialize-insecure --datadir=/dev/shm/qa/node2 --basedir=/dev/shm/qa/pxc57 2) start cluster ./pxc57/bin/mysqld --defaults-file=/dev/shm/qa/pxc57_n1.cnf --wsrep-new-cluster & ./pxc57/bin/mysqld --defaults-file=/dev/shm/qa/pxc57_n2.cnf & 3) Initiate sysbench load mysql -A -uroot -S/dev/shm/qa/node1/socket.sock -e "create user sysbench@'%' identified with mysql_native_password by 'test';" mysql -A -uroot -S/dev/shm/qa/node1/socket.sock -e "grant all on *.* to sysbench@'%';" mysql -A -uroot -S/dev/shm/qa/node1/socket.sock -e "drop database if exists sbtest;create database sbtest" sysbench /usr/share/sysbench/oltp_insert.lua --mysql-db=sbtest --mysql-user=sysbench --mysql-password=test --db-driver=mysql --mysql-socket=/dev/shm/qa/node1/socket.sock --threads=10 --tables=10 --table-size=1000 prepare sysbench /usr/share/sysbench/oltp_read_write.lua --mysql-db=sbtest --mysql-user=sysbench --mysql-password=test --db-driver=mysql --mysql-socket=/dev/shm/qa/node1/socket.sock --threads=10 --tables=10 --time=3000 --report-interval=10 run sleep 60 4) shutdown node2 for upgrade /dev/shm/qa/pxc57/bin/mysqladmin -uroot -S/dev/shm/qa/node2/socket.sock shutdown sleep 150 5) kill sysbench run # start node2 with PXC-8.0 binary ./pxc80/bin/mysqld --defaults-file=/dev/shm/qa/pxc80_n2.cnf & 6) Check error log # If above steps are not reproduced the issue, please follow these steps too to reproduce the issue. 1) shutdown cluster /dev/shm/qa/pxc80/bin/mysqladmin -uroot -S/dev/shm/qa/node2/socket.sock shutdown /dev/shm/qa/pxc80/bin/mysqladmin -uroot -S/dev/shm/qa/node1/socket.sock shutdown 2) start node1 with PXC-8.0 binary ./pxc80/bin/mysqld --defaults-file=/dev/shm/qa/pxc80_n1.cnf --wsrep-new-cluster & 3) initiate sysbench sysbench /usr/share/sysbench/oltp_read_write.lua --mysql-db=sbtest --mysql-user=sysbench --mysql-password=test --db-driver=mysql --mysql-socket=/dev/shm/qa/node1/socket.sock --threads=10 --tables=10 --time=3000 --report-interval=10 run 4) kill sysbench after 60 sec 5) start node2 ./pxc80/bin/mysqld --defaults-file=/dev/shm/qa/pxc80_n2.cnf &

Environment

None

Attachments

6
  • 04 Apr 2019, 05:31 AM
  • 04 Apr 2019, 05:31 AM
  • 04 Apr 2019, 05:31 AM
  • 04 Apr 2019, 05:31 AM
  • 04 Apr 2019, 05:31 AM
  • 04 Apr 2019, 05:31 AM

blocks

Smart Checklist

Activity

Show:

Ramesh Sivaraman April 12, 2019 at 7:20 AM

Could not reproduce the issue with the latest build.

Krunal Bauskar April 12, 2019 at 3:39 AM

I treid this issue one more time with variation and longer wait to ensure SST always. Still no success. Will mark the issue accordingly. Re-open if you re-discover this with altered test-case.

Cannot Reproduce

Details

Assignee

Reporter

Labels

Time tracking

2h logged

Affects versions

Priority

Smart Checklist

Created April 4, 2019 at 5:32 AM
Updated March 6, 2024 at 10:12 PM
Resolved April 12, 2019 at 7:21 AM