Bug 890535
Summary: | Default socket wrong for /usr/libexec/mysqld-wait-ready | ||||||
---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | John Whitley <jrw> | ||||
Component: | mysql | Assignee: | Tom Lane <tgl> | ||||
Status: | CLOSED CURRENTRELEASE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||
Severity: | unspecified | Docs Contact: | |||||
Priority: | unspecified | ||||||
Version: | 17 | CC: | hhorak, tgl | ||||
Target Milestone: | --- | ||||||
Target Release: | --- | ||||||
Hardware: | x86_64 | ||||||
OS: | Linux | ||||||
Whiteboard: | |||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | Environment: | ||||||
Last Closed: | 2013-01-12 15:17:11 UTC | Type: | Bug | ||||
Regression: | --- | Mount Type: | --- | ||||
Documentation: | --- | CRM: | |||||
Verified Versions: | Category: | --- | |||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
Cloudforms Team: | --- | Target Upstream Version: | |||||
Embargoed: | |||||||
Attachments: |
|
Thanks for reporting. (In reply to comment #0) > Initially mysql would not load because the socket nominated in /etc/my.cnf > (/tmp/mysql.sock) was not being created and so > /usr/libexec/mysqld-wait-ready did not complete. > > (Running mysqld-safe by hand did create the socket) Just to correct this -- socket probably was created, but since mysql's unit file specifies PrivateTmp, the socket was created in its own namespace. Namespace is always unique for every process, so as a consequence socket was not visible for other processes, not even for mysqld-wait-ready. > I then commented out the socket line from /etc/my.cnf to use the default. > This time the socket /var/lib/mysql/mysql.sock was created but > /usr/libexec/mysqld-wait-ready was looking for /home3/mysql/data/mysql.sock > i.e. ++ /usr/bin/mysqladmin --no-defaults > --socket=/home3/mysql/data/mysql.sock --user=UNKNOWN_MYSQL_USER ping mysqld_safe uses the following command to get the socket path: ${mysql_unix_port:-${MYSQL_UNIX_PORT:-/var/lib/mysql/mysql.sock}} so we should indeed change the following in the mysql-wait-ready: -get_mysql_option mysqld socket "$datadir/mysql.sock" +get_mysql_option mysqld socket "/var/lib/mysql/mysql.sock" Yeah, you'd have to turn off PrivateTmp if you want to keep the socket in /tmp. I think that the indirect dependency on $datadir that's now in mysqld-wait-ready was actually correct when it was written, but that was a very long time ago. I agree it's bogus now given the current coding in mysqld_safe. Will fix, thanks for the report. mysql-5.5.29-1.fc17 has been submitted as an update for Fedora 17. https://admin.fedoraproject.org/updates/mysql-5.5.29-1.fc17 mysql-5.5.29-1.fc18 has been submitted as an update for Fedora 18. https://admin.fedoraproject.org/updates/mysql-5.5.29-1.fc18 mysql-5.5.29-1.fc16 has been submitted as an update for Fedora 16. https://admin.fedoraproject.org/updates/mysql-5.5.29-1.fc16 Package mysql-5.5.29-1.fc18: * should fix your issue, * was pushed to the Fedora 18 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=updates-testing mysql-5.5.29-1.fc18' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/FEDORA-2013-0142/mysql-5.5.29-1.fc18 then log in and leave karma (feedback). mysql-5.5.29-1.fc18 has been pushed to the Fedora 18 stable repository. If problems still persist, please make note of it in this bug report. mysql-5.5.29-1.fc17 has been pushed to the Fedora 17 stable repository. If problems still persist, please make note of it in this bug report. mysql-5.5.29-1.fc16 has been pushed to the Fedora 16 stable repository. If problems still persist, please make note of it in this bug report. |
Created attachment 669635 [details] mysql /etc/my.cnf Description of problem: If 'socket' is not specified in /etc/my.cnf, /usr/libexec/mysqld-wait-ready looks for 'datadir'/mysql.sock, not /var/lib/mysql/mysql.sock Version-Release number of selected component (if applicable): mysql-server-5.5.28-2.fc17.x86_64 How reproducible: very Steps to Reproduce: 1. Install mysql 2. Use /etc/my.cnf with no socket and non-defaul datadir 3. Try to load with systemctl start mysqld.service Actual results: mysql loads but /usr/libexec/mysqld-wait-ready hangs until timeout Expected results: mysql loads Additional info: I have been using mysql 5.5.28 downloaded from oracle on pre-systemd fedora. To use an existing database on F17, I installed mysql using yum, copied /etc/my.cnf and the datadir (/home3/mysql/data) from the working installation. Initially mysql would not load because the socket nominated in /etc/my.cnf (/tmp/mysql.sock) was not being created and so /usr/libexec/mysqld-wait-ready did not complete. (Running mysqld-safe by hand did create the socket) I then commented out the socket line from /etc/my.cnf to use the default. This time the socket /var/lib/mysql/mysql.sock was created but /usr/libexec/mysqld-wait-ready was looking for /home3/mysql/data/mysql.sock i.e. ++ /usr/bin/mysqladmin --no-defaults --socket=/home3/mysql/data/mysql.sock --user=UNKNOWN_MYSQL_USER ping I then set the socket in /etc/my.cnf to /var/lib/mysql/mysql.sock and then mysql was able to be loaded by systemd and seems OK now I have added /etc/my.cnf