Server Restore to newer version

Submitted by methownet on Thu, 10/04/2012 - 02:48 Pro Licensee

I am upgrading from 8.08 to 12.04. I restored the server once but didn't have enough memory allocated and it locked up (it has about 800 mailboxes and the restore takes about an hour) so I cut bait and am trying it again. I tried the restore again over the top of the original and got this message:

Deleting old MySQL databases .. Restore failed : SQL drop database information_schema failed : Access denied for user 'root'@'localhost' to database 'information_schema'

Any idea why this might have happened and how it can be fixed ? Do I need to delete the old restoration before I try a new one?

Status: 
Active

Comments

Submitted by JamieCameron on Thu, 10/04/2012 - 14:12

That is really odd, as the information_schema database shouldn't be associated with any domain!

Try selecting the domain for which you got this error from the left menu, clicking on Edit Databases, clicking on information_schema, and then clicking the button to dis-associate it from the domain.

Then re-run the restore..

Submitted by methownet on Thu, 10/04/2012 - 19:56 Pro Licensee

the available button is "disassociate with server" under the domain. Is that the button you are talking about? I see that performance_schema is also associated with this domain. Is there any risk in disassociating these two from the server?

Submitted by methownet on Thu, 10/04/2012 - 20:19 Pro Licensee

OK. I disassociated the information_schema database from the domain/server and nothing bad seems to have happened. I will try another backup and restore at the next maintenance window and see if that does it. I noticed that when we last moved a server, that a couple of databases did not come along with the restore (they weren't associated) and my guess is that one of our web designers associated everything on the list with the server including the information database.

Submitted by andreychek on Thu, 10/04/2012 - 20:25

Yeah, it shouldn't cause any problem to disassociate that database -- it doesn't make changes to the database to do that, it just makes it so that Virtualmin doesn't try to group that database with one particular Virtual Server.

The information_schema database is a special one... it's data is read-only, and all users have access to it.

But disassociating it in Virtualmin shouldn't cause any issues.

Let us know if that helps with your imports!

Submitted by methownet on Fri, 10/05/2012 - 09:55 Pro Licensee

That did fix that problem however I now have another. The restore moves forward to this point:

Restoring SpamAssassin and Procmail configurations .. .. done

Restoring Webmin ACL files .. .. done

Restoring AWstats configuration file .. .. done

Re-creating mail and FTP users ..

and then locks there on the screen. I suspect the restore worked but how do I tell if it was successful or not? Is there a log file associated with restore? Is there a way I can tell what time the restore completed?

Submitted by JamieCameron on Fri, 10/05/2012 - 11:22

Do you mean your browser session disconnects at that point? Or just that it takes a long time (perhaps forever) before moving on to the next step?

Submitted by methownet on Fri, 10/05/2012 - 11:55 Pro Licensee

I'm not sure which. My browser session was locked on the recreating users and after 3 hours, I checked the processor load and it looks like it went back down to normal after and hour. So it appeared either to finish in an hour or lock up at the restore users part. I checked a few mailboxes and they all looked like they had current mail in them. The website looked right. But I don't know which is why I wondered if there were some way to be sure it completed the process.

Submitted by JamieCameron on Fri, 10/05/2012 - 12:06

It is possible that the restore completed, but your browser timed out.

Try going to Webmin -> Webmin Actions Log, and see if you can find a log entry for the restore.

Submitted by methownet on Tue, 10/09/2012 - 08:30 Pro Licensee

That did it. The log says it completed. Thanks. Jeff

Submitted by methownet on Mon, 10/22/2012 - 11:57 Pro Licensee

We have the new server in place now but some folks are having trouble sending mail. I see a lot of these in the log file (see below) and it seems to be related to port 995 being checked in the client. I can fix this by unchecking the ssl option for outgoing in the client, but I have 800 clients to walk through this and it would be better if it allowed both 25 and 995. Is there a way to fix this?

Oct 22 07:14:56 gto spamd[11516]: spamd: result: . 2 - HTML_MESSAGE,SPF_SOFTFAIL,T_DKIM_INVALID,URIBL_GREY scantime=0.3,size=8014,user=customername.methownet,uid=1888,required_score=8.0,rhost=localhost.localdomain,raddr=127.0.0.1,rport=35570,mid=0.0.C1.1C8.1CDB05F9A6A40AE.0@drone061.ral.icpbounce.com,autolearn=no Oct 22 07:15:47 gto spamd[11516]: spamd: result: . 1 - HTML_IMAGE_RATIO_04,HTML_MESSAGE,SPF_SOFTFAIL,T_DKIM_INVALID scantime=1.0,size=31558,user=customer.methownet,uid=1971,required_score=8.0,rhost=localhost.localdomain,raddr=127.0.0.1,rport=35595,mid=

Submitted by andreychek on Mon, 10/22/2012 - 12:32

We're happy to help you with your email issue!

Is there any chance you could open a new request for that though? That would help us keep these topics separate, as we get confused easily :-)

In your new request -- could you mention what problem your clients are seeing while sending emails? Thanks!

Submitted by BerkWood on Fri, 02/01/2013 - 17:45

I have this same issue. It is associated with the first domain on the server. I don't have a disassociate button but I have a delete