#mysql
18 October 2007
Total 25 pages. You are browsing page 5/25.
First :: Prev :: [...] [3] [4] [5] [6] [7] [...] :: Next :: Last
05:42 <****> that way you get multiple opinions
05:42 <****> Typically, the thread waiting for the binlog is in "Waiting for master to send event"
05:43 <****> ShawnRisk: http://www.tizag.com/ajaxTutorial/ajax-javascript.php
05:43 <****> okay
05:43 <****> yeah I mean the other one, typically in state "Has read all relay log; waiting for the slave I/O thread to update it"
05:43 <****> That one doesn't show up in the processlist atm
05:44 <****> fuse: weird..
05:44 <****> hmm replication broken?
05:44 <****> is that slave for failover / HA?
05:44 <****> mikegriffin: you are telling me :-)
05:44 <****> It's just for reads.
05:44 <****> WE run about 5000 SELECT queries per second during the daytime.
05:44 <****> So we have a few slaves handling all of that.
05:44 <****> ok
05:45 <****> And, according to the server and the master, it's connected and syncing.
05:45 <****> fuse: tried manually casting in the insert?
05:45 <****> and Seconds behind master?
05:45 <****> 53,041 and falling
05:46 <****> wow
05:46 <****> Also, I think that htread you're talking about is the one processing all the updates.
05:46 <****> Yeah, new box.
05:46 <****> yeah
05:46 <****> The old slave this one is replacing was synced up completely.
05:46 <****> We shutdown MySQL and rsync'd the data over to it.
05:46 <****> But, that was a couple hundred gigs and took awhile.
05:47 <****> seems normal.
05:47 <****> I suppose. I was more concerned that it wasn't being threaded.
05:47 <****> But I suppose that makes sense.
05:47 <****> yeah you'll benefit from multiple cpus/ disks when there are also reads hitting the slave
05:47 <****> but the problem with full table locks still remains
05:47 <****> mikegriffin: i shouldnt need too....
05:48 <****> Ya know, I bet the slave would sync faster if the master wasn't doing 300 updates per second right now, haha.
05:48 <****> fuse: just to rule it out
05:49 <****> fuse: like i said, kinda limited experience
05:49 <****> ok
05:49 <****> hehe
05:49 <****> mikegriffin: here is an odd thing
05:50 <****> so i am inserting a src and dst ip
05:50 <****> gfrank: which mysql version do you use?
05:50 <****> both ip's in nbo, they both print with printf fine, but only the src_ip gets inserted correct
05:50 <****> 5.0.45
05:50 <****> ok
05:50 <****> debian or ubuntu? ;)
05:51 <****> You wish
05:51 <****> centos 4.5
05:51 <****> oh well I thought because it's the most recent one for these
05:51 <****> fuse: as char?
05:52 <****> The box that we are removing can't sync at all.
05:53 <****> It's falling more and more behind.
05:55 <****> hmm you might be able to speed updates up if you run a select with the same where clause on the slave prior to the update on the master but well we are syncing currently...
05:55 <****> mikegriffin: i am printing them as ints
05:57 <****> I won't worry about it.
05:57 <****> Night's are slow for us.
05:57 <****> So long as it finishes by 9am
05:57 <****> Poor thing is only doing 212 queries per second right now
05:57 <****> fuse: ie 2130800000 == 127.1.109.128
05:57 <****> ?
05:57 <****> Curse you non-threaded replication thread!
05:58 <****> mikegriffin: i am using numeric form
05:58 <****> weird, did casting change anything?
05:58 <****> gfrank: yeah well implementing that is not trivial
05:58 <****> I know
05:59 <****> That's why I said earlier, "I suppose it makes sense."
05:59 <****> mikegriffin: i am way too tired to try right now, but i will try in the morning
05:59 <****> thankyou for the tip
06:00 <****> word
06:00 <****> finally...
06:00 <****> 2130706433 = 127.0.0.1
06:00 <****> yeah well write intensive replication with myisam and large datasets, hard thing to do ;)
06:00 <****> i was too lazy to do the math so i brute forced it
06:02 <****> whats the variable to use while starting mysql to point it to a different database location?
06:03 <****> datadir = /var/lib/mysql
06:03 <****> thanks
06:03 <****> nils_: thanks for the info
06:03 <****> no thank grep
06:03 <****> Good to know I'm not entirely wrong.
06:03 <****> gfrank: well I doubt that was new for you
06:03 <****> Just rusty on my terminology.
06:03 <****> well im out gotta go to my 2130706433
06:04 <****> (ping it :)
06:04 <****> Really, I was hoping there was something I was missing, which is why I came here.
06:04 <****> where would the dbs be stored for mysql on linux?
06:05 <****> ChrisC35: in the datadir;
06:05 <****> where is that by default?
06:06 <****> ChrisC35: show variables like 'datadir';
06:06 <****> i cant
06:06 <****> im looking throu a hd in rescue mode in linux
06:06 <****> trying to grab my dbs
06:06 <****> which OS?
06:06 <****> centos
06:07 <****> hmm don't know
06:08 <****> should be somewhere in /var/lib or /var
06:08 <****> It's rather intimidating when you're spending lots of money on colo equipment, especially for something with these demands ...
06:08 <****> DO hope that SATA will suffice.
06:08 <****> To get enough disk space to maintain at least a year's growth, we'd be spending a niiiiice chunk of change on scsi.
06:09 <****> ChrisC35: /var/lib/mysql on Centos 5 yum install
06:09 <****> thanks
06:09 <****> hi folks
06:09 <****> yeah you might consider WD Raptor Disks...
06:10 <****> nils_++
Total 25 pages. You are browsing page 5/25.
First :: Prev :: [...] [3] [4] [5] [6] [7] [...] :: Next :: Last