11 messages in com.mysql.lists.bugsRe: Diagnostic route for this replica...| From | Sent On | Attachments |
|---|---|---|
| Whit Blauvelt | 13 Jun 2001 12:49 | |
| Sasha Pachev | 14 Jun 2001 10:01 | |
| Whit Blauvelt | 14 Jun 2001 11:31 | |
| Sasha Pachev | 14 Jun 2001 12:19 | |
| Whit Blauvelt | 24 Jun 2001 22:42 | |
| Michael Widenius | 25 Jun 2001 00:28 | |
| Whit Blauvelt | 25 Jun 2001 17:55 | |
| Sasha Pachev | 25 Jun 2001 19:37 | |
| Whit Blauvelt | 26 Jun 2001 10:36 | |
| Whit Blauvelt | 26 Jun 2001 11:05 | |
| Sasha Pachev | 26 Jun 2001 11:20 |  .pub |
| Subject: | Re: Diagnostic route for this replication problem? |
|---|---|
| From: | Whit Blauvelt (wh...@transpect.com) |
| Date: | 06/26/2001 11:05:44 AM |
| List: | com.mysql.lists.bugs |
On Mon, Jun 25, 2001 at 08:37:51PM -0600, Sasha Pachev wrote:
CHANGE MASTER TO MASTER_HOST='china.patternbook.com',MASTER_PORT=3307,MASTER_PASSWORD='password_on_the_master', MASTER_USER='ftp';
Sasha,
There is one other possibility that just occurred to me about where the bug could be. It used to be that the replication mini-client could connect for replication via the stunneled port 3307 even if it the slave was pointed at 'localhost'. Something was changed since .27 so that now that doesn't work - it needs the local machine to be specified by name just like the mysql client. Loading the table from the master definitely is connecting properly to 3307, but perhaps the replication mini-client, even when the host name is specified rather than localhost, is connecting to port 3306 instead, and getting the slave rather than the master?
No, that doesn't seem to be it - setting the same privileges on the slave for the user doesn't result in a successful login either, still:
010626 12:58:31 Slave: Failed reading log event, reconnecting to retry, log
'FIRST' position 4
010626 12:58:31 Slave: reconnected to master
'ft...@china.patternbook.com:3307',replication resumed in log 'FIRST' at position
4
010626 12:58:31 Error reading packet from server: Access denied for user:
'ftp@localhost' (Using password: YES) (read_errno 0,server_errno=1045)
Oh well, at least that rules that out. Unless the fact that the slave is not actually logging as a master would also produce the "Access denied" error?
BTW, master.info looks like:
[blank line] 4 china.patternbook.com ftp password_on_the_master 3307 60
Whit