Bug #47366 | ndb_config --xml --configinfo: show restart types | ||
---|---|---|---|
Submitted: | 16 Sep 2009 11:53 | Modified: | 28 Dec 2009 19:48 |
Reporter: | Sivert Sørumgård | Email Updates: | |
Status: | Closed | Impact on me: | |
Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S3 (Non-critical) |
Version: | mysql-5.1-telco-7.0 | OS: | Any |
Assigned to: | Magnus Blåudd | CPU Architecture: | Any |
Tags: | 7.0.6 |
[16 Sep 2009 11:53]
Sivert Sørumgård
[17 Sep 2009 5:57]
Sveta Smirnova
Thank you for the reasonable feature request.
[29 Oct 2009 12:24]
Magnus Blåudd
Will make it possible to flag each parameter with the below suggested flags and then add the resulting tags in the .xml file based in that. /* Flags telling how the system must be restarted for a changed parameter to take effect Default is none of these flags set, which means node restart of one node at a time for the setting to take effect CS_RESTART_INITIAL Each data node need to be restarted one at a time with --initial CS_RESTART_SYSTEM The whole system need to be stopped and then started up again CS_RESTART_SYSTEM + CS_RESTART_INITIAL The whole system need to be stopped and then restarted with --initial thus destroying any data in the cluster */ CI_RESTART_SYSTEM = 4, // System restart is necessary to apply setting CI_RESTART_INITIAL = 8 // Initial restart is necessary to apply setting
[3 Nov 2009 14:48]
Bugs System
A patch for this bug has been committed. After review, it may be pushed to the relevant source trees for release in the next version. You can access the patch from: http://lists.mysql.com/commits/89148
[3 Nov 2009 15:33]
Magnus Blåudd
Approved on IRC and phone
[3 Nov 2009 16:54]
Magnus Blåudd
Pushed to 6.3 and 7.0.10
[4 Nov 2009 14:57]
Jon Stephens
Documented bugfix in the NDB-6.3.29 and 7.0.10 changelogs as follows: The output from ndb_config --xml --configinfo now includes the restart type required for resetting each configuration parameter. Closed.
[17 Dec 2009 19:37]
Jon Stephens
The restart info now appears in ConfigInfo.cpp, but is not present in the output of 'ndb_config --configinfo' or 'ndb_config --configinfo --xml'. Users should not have to read source files to get this info. Checked with 6.3.29/7.0.10/7.1.0 built yesterday. See also BUG#49781.
[18 Dec 2009 21:15]
Magnus Blåudd
The restart type is not printed if not changed from default. I.e all parameters that does not say anything about restart uses the default value, which is to restart ALL nodes in the cluster(ndb_mgmd, ndbd and api nodes). Currently only very limited number of parameters have a value that differs from the default. See my "grep" below: msvensson@pilot:~/mysql/7.0/storage/ndb/tools$ ./ndb_config --configinfo --xml | grep restart <param name="NoOfReplicas" comment="Number of copies of all data in the database (1-4)" type="unsigned" default="2" min="1" max="4" restart="system" initial="true"/> <param name="StopOnError" comment="If set to N, ndbd(DB) automatically restarts/recovers in case of node failure" type="bool" default="true"/> <param name="Diskless" comment="Run wo/ disk" type="bool" default="false" restart="system" initial="true"/> <param name="LogLevelNodeRestart" comment="Node restart, node failure info printed on stdout" type="unsigned" default="0" min="0" max="15"/> <param name="DiskCheckpointSpeedInRestart" comment="Bytes per second allowed to be written by checkpoint during restart" type="unsigned" default="104857600" min="1048576" max="4294967039"/> <param name="InitialLogfileGroup" comment="Logfile group that will be created during initial start" type="string" restart="system" initial="true"/> <param name="InitialTablespace" comment="Tablespace that will be created during initial start" type="string" restart="system" initial="true"/> <param name="BuildIndexThreads" comment="No of threads to use for building ordered indexes during system/node restart" type="unsigned" default="0" min="0" max="128"/> <param name="NoOfReplicas" comment="Number of copies of all data in the database (1-4)" type="unsigned" default="2" min="1" max="4" restart="system" initial="true"/> <param name="StopOnError" comment="If set to N, ndbd(DB) automatically restarts/recovers in case of node failure" type="bool" default="true"/> <param name="Diskless" comment="Run wo/ disk" type="bool" default="false" restart="system" initial="true"/> <param name="LogLevelNodeRestart" comment="Node restart, node failure info printed on stdout" type="unsigned" default="0" min="0" max="15"/> <param name="DiskCheckpointSpeedInRestart" comment="Bytes per second allowed to be written by checkpoint during restart" type="unsigned" default="104857600" min="1048576" max="4294967039"/> <param name="InitialLogfileGroup" comment="Logfile group that will be created during initial start" type="string" restart="system" initial="true"/> <param name="InitialTablespace" comment="Tablespace that will be created during initial start" type="string" restart="system" initial="true"/> <param name="BuildIndexThreads" comment="No of threads to use for building ordered indexes during system/node restart" type="unsigned" default="0" min="0" max="128"/>
[19 Dec 2009 11:38]
Jon Stephens
Understood, I'll update changelog entry/ndb_config description and my scripts that pull this info out of XML output for docs to accomodate this. Thanks for the clarification.
[28 Dec 2009 19:17]
Jon Stephens
Updated changelog entry reads, The output from ndb_config --configinfo --xml now indicates, for each configuration parameter, whether nodes need to be restarted using the --initial option when resetting that parameter. Also updated ndb_config --xml option description in Manual. Closed.
[29 Dec 2009 9:38]
Jon Stephens
Recent email discussion with Magnus has to revealed to me that I've been imbibing too much or too little Yuletide glögg, and that I just wasn't 'getting it'. I've updated the changelog entry and related info in the description of ndb_config's --xml option. The docs commit is here: http://lists.mysql.com/commits/95775