Apparently nothing. Just the "Show process list" query.
70100 | <user> | localhost | <database> | Query | 0 | NULL |
show full processlist
+1 612 859 8161
On 01/03/2011 01:51 AM, Ananda Kumar wrote:
> What is the output of "show full processlist"
> On Mon, Jan 3, 2011 at 1:10 PM, Sairam Krishnamurthy
> <kmsram420@stripped <mailto:kmsram420@stripped>> wrote:
> Hi all,
> I set up replication some days back. Because of some internal
> issues we canceled the replication and reset the master to be a
> normal database server. But after this "revert", the database is
> unusually slow.
> It was slow because of the discussion below when replication was
> on. But I wonder why it is slow now after the master has been
> reset. By reset I mean, commenting out the master setup lines in
> the my.cnf file, removal of logs and restarting the database.
> Thanks in advance
> Sairam Krishnamurthy
> +1 612 859 8161
> On 12/06/2010 04:47 AM, Ananda Kumar wrote:
> Also, make sure your /tmp folder is on a separate and fast disk.
> We had similar issues and we moved /tmp folder from Local to
> SAN storage and it was quite fast.
> On Mon, Dec 6, 2010 at 4:10 PM, Johan De Meersman
> <vegivamp@stripped <mailto:vegivamp@stripped>
> <mailto:vegivamp@stripped <mailto:vegivamp@stripped>>>
> Are you saying that mass inserts go much slower now that
> you've set up
> replication? In that case, I suspect you have your binlogs
> on the
> same disk
> as your data.
> Put the binary logs on separate disks, and you'll notice a
> dramatic increase
> in performance.
> On Mon, Dec 6, 2010 at 7:17 AM, Sairam Krishnamurthy
> <kmsram420@stripped <mailto:kmsram420@stripped>
> > All,
> > I have a situation. We have a very large database(gigabytes).
> When we load
> > the data in to the table, it will be huge again (20 million
> rows). So in
> > order to get good backup we set up a replication(master-slave
> setup). But
> > now, because of this setup, loading into the table takes hours
> which will be
> > generally done in a few minutes otherwise.
> > Is there a workaround for this? We really don't want to wait for
> hours for
> > the table to be loaded. Should we abandon replication for this?
> Any pointers
> > ?
> > --
> > Thanks,
> > Sairam Krishnamurthy
> > +1 612 859 8161
> Bier met grenadyn
> Is als mosterd by den wyn
> Sy die't drinkt, is eene kwezel
> Hy die't drinkt, is ras een ezel