Scheduled backups fail to run

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Create a new instance, under D0 plan
2. Schedule backups
3. Note that backups fail to run

What is the name of your database instance?

crowdrankersql:crowdrankerinst1

What is the expected output? What do you see instead?

I have scheduled backups to run at 10am UTC, and for several days, they have 
not run yet. 

Please provide any additional information below.

I have also tried disabling, then re-enabling backups, but all I see is an 
information stating that "This backup has not run yet", with no indication on 
the reason.

Original issue reported on code.google.com by luca.de.alfaro@gmail.com on 31 Mar 2013 at 5:52

[GoogleCodeExporter]

In case this helps, I am connecting to Google Cloud SQL from web2py, using a 
DAL connection string. 
All works fine.  I did no change to the user table, but if I do 

select * from mysql.user;

I do not see the cloudsqladmin user listed. 
I am unsure though of why web2py would actually delete such user -- I will ask 
the web2py community on this, but I somehow highly doubt that web2py would drop 
users from a database.  Is it possible that such user was never created?  How 
does one restore it? 

Original comment by luca.de.alfaro@gmail.com on 31 Mar 2013 at 6:18

[GoogleCodeExporter]

For what it is worth, I created a new instance at level D1 (which I have since 
deleted, as I was only testing).  Even in this instance, when I execute "select 
* from mysql.user" I do not see the cloudsqladmin user listed.  All I see is 
four rows, three for user root at localhost, 127.0.0.1, and ::1, and one for 
localhost. 

Original comment by luca.de.alfaro@gmail.com on 31 Mar 2013 at 6:44

[GoogleCodeExporter]

The cloudsqladmin user is expected to not show up in mysql.user. The fact that 
the backups did not run is indeed a bug. We recently fixed it and the fix will 
be deployed in prod later this week.

Original comment by razv...@google.com on 1 Apr 2013 at 4:52

• Changed state: PendingRelease

[GoogleCodeExporter]

I can confirm that the backup has run tonight.  Wonderful and thank you for the 
very quick reply. 
Let me also just add how much I like the service!  A big thanks to you all. 

Original comment by luca.de.alfaro@gmail.com on 1 Apr 2013 at 2:38

[GoogleCodeExporter]

Backups are not occurring daily.  This is what I see for my most recent ones: 
Apr 6, 2013 3:01 AM     
IBacked up volume
Apr 4, 2013 3:12 AM     
IBacked up volume

April 5 and 7 had no attempts to have a backup. 

Original comment by luca.de.alfaro@gmail.com on 7 Apr 2013 at 9:41

[GoogleCodeExporter]

If there is no activity on your instance since the previous backup, then the 
daily backup will be skipped. Could you confirm if that is the case?

Original comment by t...@google.com on 8 Apr 2013 at 5:22

• Changed state: New

[GoogleCodeExporter]

I am sure that at least on April 7 there was activity.  Not a lot of it -- my 
site is just new -- but there were some db writes.

Original comment by luca.de.alfaro@gmail.com on 8 Apr 2013 at 5:45

[GoogleCodeExporter]

You likely hit a product bug that we have fixed and are deploying in prod this 
week. I can confirm that the backup has run for April 8. 

Original comment by deban...@google.com on 8 Apr 2013 at 8:00

[GoogleCodeExporter]

Please let us know if you notice missing backups again.
Please note that the scheduled time is in UTC, and the backup can occur within 
a 4 hour window after the scheduled time.

Original comment by t...@google.com on 10 Apr 2013 at 12:51

• Changed state: Fixed