Re: cloudSQL's mysterious action.

- Output of "SHOW PROCESSLIST" when the problem happens
----------------------------------------------------------------
Id User Host db Command Time State Info
249283 root localhost runwaychannel Query 0 NULL SHOW PROCESSLIST
----------------------------------------------------------------


- Output of "SHOW ENGINE INNODB STATUS" when the problem happens
----------------------------------------------------------------
Type Name Status
InnoDB ===================================== 121006 7:38:18 INNODB MONITOR OUTPUT ===================================== Per second averages calculated from the last 1 seconds ----------------- BACKGROUND THREAD ----------------- srv_master_thread loops: 194225 1_second, 194163 sleeps, 18769 10_second, 8915 background, 8910 flush srv_master_thread log flush and writes: 201654 ---------- SEMAPHORES ---------- OS WAIT ARRAY INFO: reservation count 1536320, signal count 1262054 Mutex spin waits 155556036
----------------------------------------------------------------

Thank you for your reply, and sorry for not enough informations.

I can show you additional informations, though there are some which cannot 

be shown

on this public board.

>- Exact time of day with timezone

Happened several times.

Latest one was occurred in 9:34, 3 Oct.(UTC)

>- Exact error messages

Does this mean which error messages?

On the side of applications, "cannot connect to database" was the message.

>- Instance name

I cannot show this since the name of our customer is included in the 

instance name.

If it is possible by contacting over email, I can tell my email address to 

you on this

board. Please let me know if you can contact us by email or not.

>- Output of "SHOW PROCESSLIST" when the problem happens

>- Output of "SHOW ENGINE INNODB STATUS" when the problem happens

There are no evidences of above at the time when the error takes place.

But I attach a file which includes above, in normal situation.

Sorry for not sufficient informations, but thank you for your help.

And I need your help ongoing.

2012年10月5日金曜日 22時07分04秒 UTC+9 西内功:

Let me ask a quetion about cloudSQL which are carrying out a mysterious action.

Although the site is operated by the unrestricted plan of cloudSQL, there occurs a situation with no response from DB.

It has become impossible to use the site with a connection error of DB which is shown on the site.

At the time above occurs, when I check with "Google APIs Console", 

connection becomes 0 and it seems that the DB has been rebooted somehow. 

After 10 to 15 minutes when connection becomes 0, the number of connections is increased at a stroke as the queries which had accumulated are processed.

Although it has happened several times, it occurs not at the time especially with many accesses, rather occurs at time with few accesses, such as 

midnight. It seems that this is not a problem of applications because of the frequency of this happening.

Does this phenomenon occur by the structure of applications?

Or is it the specification of GAE?