8.01 latest update FTP backup returns - but doesn't work :(

[RuthServer]

"at the end of august"??????? a joke????!!!!! this is a critical Bug.. i want a fix now!!!

 

[PE-Steve]

My server has seemed to stop giving me the errors about the FTP path, etc. And yes, I'm not too fond of waiting until the end of August either, but I don't mind if they're actually going to test the update BEFORE they release it. We all know how well SWSoft's products work when the first come out. Oh wait, they don't.

 

[0031]

To disable scheduled backups at all, comment out or erase crontab record, running backupmng every 15 minutes.

Can someone tell me where all this redundant 15-min backups are stored? In local repository or psa/dumps/tmp dir? Can you provide me with listing of it?

And can I ask you to show me database structure for those domains, which produce "Couldn't resolve host 'ftp:'" message? In particular I'm interesting in:

# mysql -uadmin -p`cat /etc/psa/.psa.shadow` psa

mysql> select id from domains where name="the problem domain name here";

mysql> select * from BackupsScheduled where obj_id=DOMAIN_ID_HERE and obj_type='domain';

mysql> select * from BackupsSettings where id=DOMAIN_ID_HERE and type='domain';

Anyone please?

 

[0031]

To remove "Couldn't resolve host 'ftp:'" error:

# mysql -uadmin -p`cat /etc/psa/.psa.shadow` psa -e 'update BackupsScheduled set repository="local";'

 

[0031]

Space consumption is a consequence of FTP upload failure: Backup Manager didn't erase backup file, if it can't be uploaded to FTP location, and retries to backup domain again on the next run.

 

[k.satirli]

my backups are stored in:

/var/lib/psa/dumps/<clientid>/<domainid>

 

[0031]

Guys, can somebody tell me the database structure for problem domains, or let me know that my proposed solution doesn't work and I should come one's way with my stupid advices?

- Doctor, everyone ignores me...
- Next please!

 

[PE-Steve]

Originally posted by 0031
To remove "Couldn't resolve host 'ftp:'" error:

# mysql -uadmin -p`cat /etc/psa/.psa.shadow` psa -e 'update BackupsScheduled set repository="local";'

Does that really work, or is that what you're posting about now?

 

[0031]

Originally posted by PE-Steve
Does that really work, or is that what you're posting about now?

As I said, I can't repeat this problem on my live server. I obtained this error message only when I manually registered 'ftp' as a backup storage in psa.BackupsScheduled database table, without appropriate records in psa.BackupsSettings (with login/host/password/etc)

So, I assume, to remove that error it is sufficient to explicitly register 'local' backup storage for all scheduled backups. It's exactly what 'update BackupsScheduled set repository="local";' do...

And I also want to see results of:

> select * from BackupsScheduled where obj_id=DOMAIN_ID_HERE and obj_type='domain';
> select * from BackupsSettings where id=DOMAIN_ID_HERE and type='domain';

to determine if my BackupsScheduled.repository assumption was right or wrong...

 

[Musicman75]

The same problem as I have. Additionally I got the following message when clicking on the backup button (domain and client):

Unable get daemon current status: Unable to create UNIX socket /usr/local/psa/tmp/agent-runner.31090: Connection refused

Is it possible to downgrade to version 8.0?

 

[Pagemakers]

Please please please, how do we stop these fuc**ng backups?

I hit the OK button and then a day or so later they are all doing it again.

This is possibly the biggest Plesk fuc* up to date.

 

[Pagemakers]

Does this fuc*up also fill up anything else in the var folder?

I have emptied all th backups and still it is fuller than it was.


EDIT: Found out that it was a master backup of all the erroneous backup files!

 

[madadam]

In case it's any help to anyone else...

I managed to stop the problem by setting-up an FTP account for the client with the backup, and setting the backup to use FTP (otherwise the option is greyed out). Then switch back to local (leaving the unused FTP settings), and it seems to run OK (been OK for over 24 hours now, so fingers crossed)...

 

[Pagemakers]

Plesk say they have foxed my server by updating a few RPM files. I will monitor the situation for a day or so and then I think they are going to release another update to fix this woeful mess.

 

[Pagemakers]

Ok. The update Plesk installed on my servers has fixed many of the problems, however there is still one big problem. On the scheduled backup page I have limited the number of backups to 1 but this is still not happening and multiple backup files are being created.

Reported to Plesk.

 

[Pagemakers]

Backup now works correctly. I'm sure there will be a update soon to fix this issue.

 

[PE-Steve]

Did you/your Plesk provider just open a ticket with SWSoft? I'd sure like to get this fixed, although I don't want to take to chances of their RPM update messing up the next Plesk update.

 

[Pagemakers]

I opened a ticket with them personally.

 

[Pagemakers]

Well sadly, this problem is not fixed at all.

Now, when a domain makes a backup to the remote FTP server it deletes EVERY other backup on there regardless of the domain. Do a backup of domain1.com would delete all backups from domain2.com, domain3.com etc.

I had 25 backups on each FTP backup server, now I only have 1 from the latest Plesk backup.

Another poorly tested release.

 

[Pagemakers]

Pleased to say this looks like it has finally been resolved with the files PLesk have just installed to my servers.