tag:blogger.com,1999:blog-19281624.post4250526278007814393..comments2023-11-29T09:00:11.662+00:00Comments on Johan Andersson's Cluster and HA Blog: MySQL Cluster: Disk Data ConfigJohan Anderssonhttp://www.blogger.com/profile/09202425156709885026noreply@blogger.comBlogger6125tag:blogger.com,1999:blog-19281624.post-23298337929609094512012-08-16T16:43:08.143+01:002012-08-16T16:43:08.143+01:00Hi,
I have 2 hard drives on each data node. (1TB ...Hi,<br /><br />I have 2 hard drives on each data node. (1TB HDD and 480GB SSD). How do I have this (A), (B) and (C) combined? What data must be stored on SSD, so I effectively use SSD? On which hard drive I need to install system and cluster? <br />'.dat' => 225GB<br />'.log' => 84GB<br />'LCP and D1 ..' => 182GBburaknoreply@blogger.comtag:blogger.com,1999:blog-19281624.post-68544662665014999602012-08-16T16:41:05.881+01:002012-08-16T16:41:05.881+01:00Hi,
I have 2 hard drives on each data node. (1TB ...Hi,<br /><br />I have 2 hard drives on each data node. (1TB HDD and 480GB SSD). How do I have this (A), (B) and (C) combined? What data must be stored on SSD, so I effectively use SSD? On which hard drive I need to install system and cluster? <br />'.dat' => 225GB<br />'.log' => 84GB<br />'LCP and D1 ..' => 182GBBuraknoreply@blogger.comtag:blogger.com,1999:blog-19281624.post-73091307381528137022012-08-14T00:13:44.930+01:002012-08-14T00:13:44.930+01:00Hi,
I have never seen or heard about this. What I...Hi,<br />I have never seen or heard about this. What I can tell you have not missed anything, but i think you are just unlucky and have hit a bug.<br /><br />Have you opened a bug report?<br /><br />-JJohan Anderssonhttps://www.blogger.com/profile/09202425156709885026noreply@blogger.comtag:blogger.com,1999:blog-19281624.post-20973417779822764032012-08-13T15:55:14.124+01:002012-08-13T15:55:14.124+01:00Hey Johan,
I ran into a weird problem in 7.2.7 th...Hey Johan,<br /><br />I ran into a weird problem in 7.2.7 that I was curious if you had seen. For testing, I installed on two Intel E5-1650s, and started it up. With only 12 cores, I gave it 4 LQHs, 1 send, 1 recv, 1 rep, 2 TCs, and had the main and IO running together; leaving two cores for the OS. When I tried to do a backup, after writing the schemas to the CTL file, the first LQH thread tried to send a GSN_SCAN_FRAGREQ (353) signal to the second LQH thread. Since they aren't supposed to communicate, no JB exists and so the data nodes receive a SIGSEGV from the OS. <br /><br />Did I miss some critical bit of info in implementing ThreadConfig? Jay Wardhttps://www.blogger.com/profile/08865692136540517839noreply@blogger.comtag:blogger.com,1999:blog-19281624.post-12677371521871406692012-05-16T14:24:32.048+01:002012-05-16T14:24:32.048+01:00Hi Parry,
Crash recovery is automatic.
Basically ...Hi Parry,<br /><br />Crash recovery is automatic.<br />Basically tables are checkpointed to disk (or table space for disk data tables), and there is a redo log written too.<br /><br />Those components are used during both node recovery (one or more nodes down, but cluster is still up) and system recovery (all nodes are dead, cluster dead), as well as automatically copying changes from one node with more data to the recovering node.<br /><br />Then you can use binary logs to do point in time recovery, because since the data nodes syncs the redo log to disk every TimeBetweenGlobalCheckpoints then it is a chance to loose that amount of data IF ALL nodes in the cluster crash. <br /><br />But during normal operation if you have a node failure, you will not lose any data. <br /><br />Hope this helps somewhat,<br />JohanJohan Anderssonhttps://www.blogger.com/profile/09202425156709885026noreply@blogger.comtag:blogger.com,1999:blog-19281624.post-86340585117368641102012-05-16T14:13:37.727+01:002012-05-16T14:13:37.727+01:00Hi Johan,
I am pretty new to MySQL Cluster. Our c...Hi Johan,<br /><br />I am pretty new to MySQL Cluster. Our company is planning to use MySQL cluster for its new product.<br /><br />I have went throught the backup and recovery options but could not find "crash recovery" anywhere i.e., recovering cluster db from a non-consistent state.<br /><br />I also read that binary logs are not transaction logs and can't be used for crash recovery.<br /><br />Is crash recovery possible in MySQL cluster and if it is possible, can you pls suggest how to do it or let me know a resource where I can study it?<br /><br />I'll appreciate your reply. Have a great day..!!Parrynoreply@blogger.com