Occasionally I will notice that one of my slaves is not in sync. It will be reading fine from the master's bin-log, but it gets hung up in the relay log and stops. It seems there is corrupt data in the relay log. When I use the mysqlbinlog utility, I get tons of "mysqlbinlog: Error writing file 'UNOPENED' (Errcode: 29)", and when it does show the queries, the queries are all messed up. Missing the first half of a big select or insert - so when it gets to one of those in the relay-log it also quits.
This causes a great problem because I cannot "fix" this issue by pointing it ahead in the relay log. Once this happens the first time, the rest of the log is riddled with them. I may not notice this for a day or more. The master gets TONS of writes, it is very highly used. I'm wondering if this could be something where the disks themselves are too slow for the application? Is that possible? Or should it just queue everything it needs to write? Or is there a buffer somewhere for this? So I have two questions: a) What is going on here, can I fix this easily (or not so easily even!) b) Is there a way to have mysql easily notify me via email, or hitting a url or some sort of automagical device to tell me things are stopped, or is the best way to write a quick script to pull show slave status and check the vars? Thank you, Adam ########################################### This message has been scanned by F-Secure Anti-Virus for Microsoft Exchange. For more information, connect to http://www.F-Secure.com/ -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
