2012/10/25 Sabika M
> I have replication setup in the following way:
>
> A -> B ->C
>
> I am making updates to server A. I want to stop all my updates and point
> them to server C. After I start writing to server C, can I use the change
> master statement to make the C the master of A (take B out
A guess -- MEM is issuing that on a timed basis, but the query is running
longer than the time period. Can you slow MEM down?
Meanwhile, file a bug report. You paid too much money for this kind of junior
mistake.
From: Sabika M [mailto:sabika.makhd...@gmail.com]
Sent: Thursday, October 25, 201
Thank you Rick! You're a super freak! ;-p
This gets me in the vicinity of where I'm trying to go and I learned a few
new tricks with the StartLoop: stuff too! Neat!
d.
> -Original Message-
> From: Rick James [mailto:rja...@yahoo-inc.com]
> Sent: Thursday, October 25, 2012 1:54 PM
> To: D
I don't have the full anymore. It had a few hundred of these. I had issues
a drop database that hung which made me run show processlist. Sadly, I did
not check the processlist beforehand.
| 101189 | monitor | localhost:42585 | NULL | Query | 4079 |
NULL |
What does SHOW FULL PROCESSLIST say?
What values do you have for
max_connections
wait_timeout (GLOBAL version)
> -Original Message-
> From: Sabika M [mailto:sabika.makhd...@gmail.com]
> Sent: Thursday, October 25, 2012 11:44 AM
> To: mysql@lists.mysql.com
> Subject: MySQL Monitor and Perc
If the 90 days is back from MAX(created_on) for a given customer...
INDEX(customer_id, created_on)
will probably be needed. And that should replace KEY `customers_id`
(`customer_id`).
Maybe...
DELETE FROM customers_log AS a
WHERE a.customer_log_id >= @left_off AND a.customer_log_id < @z
Well, the customer_id is relevant in that I want the last 90 days relative
to each customer.
customer_id = 123 might have logs from jan - mar
customer_id = 444 might have logs from feb - may
So it's a rolling log of THEIR last 90 days from their last log (most
recent) back 90 days from there.
Do
Sound right.
Be sure there are no writes during certain critical times. And that
replication is caught up.
If you want to verify the "sameness" see Percona's pt-table-checksum.
> -Original Message-
> From: Sabika M [mailto:sabika.makhd...@gmail.com]
> Sent: Thursday, October 25, 2012 1
Off hand, I would iterate over the PRIMARY KEY, looking at a thousand rows at a
time, DELETEing any that need to be purged. I would use a Perl or PHP loop, or
write a stored procedure. More discussion of "huge deletes" (which this
_could_ be):
http://mysql.rjweb.org/doc.php/deletebig
(PART
I have a customer log table that is starting to rapidly fill up (we have
hundreds of thousands of users, but many are transient, and use the service
for a few months, or use the free trial and quit, etc.)
CREATE TABLE `customers_log` (
`customer_log_id` bigint(20) unsigned NOT NULL auto_incremen
We are in the process of switching to percona binaries of MySQL . I am
using Percona 5.5.27 and monitoring the MySQL server with the MySQL
Enterprise monitor. It starts up fine, but after a while I end up with
MySQL monitor connections stacking up until the server becomes pretty much
useless. This
Dehua,
Thanks. You are correct. Perhaps I was just over-thinking it.
-Hank
On Fri, Oct 19, 2012 at 9:48 PM, Dehua Yang wrote:
> Hi Hank
>
> I just can think like this table
>
> CREATE TABLE xyz (
> hashtag VARCHAR(...) NOT NULL,
> comment_id ... NOT NULL,
> user_id bigint unsigned not null,
>
I have replication setup in the following way:
A -> B ->C
I am making updates to server A. I want to stop all my updates and point
them to server C. After I start writing to server C, can I use the change
master statement to make the C the master of A (take B out of the topology)
and proceed to s
Not sure how the .sql backup file would have done that, but it seems to
be working now. I'll just step away slowly and carry on. Thanks for
the help!
On 10/25/2012 2:45 AM, Nitin Mehta wrote:
Hi Tim,
It looks like your '.sql backup file' has changed the password for root user
and why it is
14 matches
Mail list logo