If you replicate a long command; more specifically, a command (or transaction) that started a long time ago, it will appear that replication is far behind. This is because the apparent delay is computed [I think] from the diff of the current time on slave and the time that the currently replicated command _started_ on the _master_.
Granted, 210000 sounds excessive. > -----Original Message----- > From: Martijn van den Burg [mailto:[EMAIL PROTECTED] > Sent: Friday, May 19, 2006 8:35 AM > To: sheeri kritzer > Cc: mysql@lists.mysql.com; [EMAIL PROTECTED] > Subject: Re: Momentary huge replication lag > > Hi, > > On Friday 19 May 2006 17:12, sheeri kritzer wrote: > > I've found that queries that take a long time cause lag time. > > Yes, I know, especially on a busy server. The master handles multiple > statements (connections) in parallel, but the slave processes > them serially. > > My 4.1.10 setup does about 170 queries/second, of which about > a third are > update statements, and I hardly ever got a lag above > approximately 900 > seconds. > > > Replication on a slave has 2 threads -- one to retrieve > stuff from the > > logs, and another to actually run the DML queries. Therefore, while > > one thread is stuck on a looooonnnnggg query, the other thread is > > still gathering stuff from the master, and that causes lag times. > > > > The value you gave for seconds_behind_master is about 58 > hours -- that > > seems unusually high. What are you doing to flush logs, etc? > > Nothing special, just some updates on a single database. No > flushing of > logs... The strange thing is that the condition of extremely > high lag lasts > only a couple of seconds, and then tapers back very quickly to zero. > > Martijn > > > > On 5/19/06, Martijn van den Burg <[EMAIL PROTECTED]> wrote: > > > Hi, > > > > > > Recently I created a new replication set up with > 5.0.18-standard-log on > > > Solaris 8 (one master, one slave). > > > > > > Replication is running, but periodically (after a bunch > of INSERT/UPDATE > > > statements) there is a very large replication lag, with > > > Seconds_Behind_Master values of 210000 and more. This > situation lasts for > > > a few seconds and then all is normal. > > > > > > What could be the cause? I never had this happen with 4.1.10. > > > > > > Apologies for cross-posting; the volume on the > 'replication' list is so > > > low that I feared there might be no answer ;) > > > > > > > > > Regards, > > > > > > Martijn > > > > > > -- > > > MySQL General Mailing List > > > For list archives: http://lists.mysql.com/mysql > > > To unsubscribe: > http://lists.mysql.com/[EMAIL PROTECTED] > > -- > </Martijn> > > -- > MySQL Replication Mailing List > For list archives: http://lists.mysql.com/replication > To unsubscribe: > http://lists.mysql.com/[EMAIL PROTECTED] > > -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
