[Plesk 12.5 CentOS 7] Migration fails no matter what

[Boindil]

Hi,
I'm trying to migrate to a new server. However the migration always fails. I resolved all pre-migration errors/warnings, but its still failing.

+|2015-11-10_15:14:50,021|D|MT|core.utils.common.http_xml_client|||API request to https://IP:8443/enterprise/control/agent.php:
=|2015-11-10_15:14:50,021|D|MT|core.utils.common.http_xml_client|||<?xml version='1.0' encoding='utf-8'?>
=|2015-11-10_15:14:50,021|D|MT|core.utils.common.http_xml_client|||<packet version="1.6.6.0"><webspace><get><filter><name>domainname</name></filter><dataset><gen_info /></dataset></get></webspace></packet>
+|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||API response from https://IP:8443/enterprise/control/agent.php:
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||<?xml version='1.0' encoding='utf-8'?>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||<packet version="1.6.6.0">
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||  <webspace>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||    <get>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||      <result>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||        <status>error</status>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||        <errcode>1013</errcode>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||        <errtext>domain does not exist</errtext>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||        <filter-id>domainname</filter-id>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||      </result>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||    </get>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||  </webspace>
=|2015-11-10_15:14:50,146|D|MT|core.utils.common.http_xml_client|||</packet>

=|2015-11-10_15:14:51,780|D|MT|core.workflow.runner.by_subscription|||MigrationError: Unable to continue migration until there are no issues at 'error' level in pre-migration checks.Please review pre-migration tree above and fix the errors. You can also use --ignore-pre-migration-errors command- line option if you're sure what you are doing.

I also tried a manual backup restore. The backup file has been encrypted with a custom password and is 50GB in size. Thus I cannot upload it to the server via webinterface since it says "Request entitiy too large".

I transferred the file to the new server via SFTP and tried to restore it via

/usr/local/psa/bin/pleskrestore --restore /root/backup_1.tar -level clients -backup-password XXX -ignore-sign

I'm getting this output

Unable to import file as dump: The file you are trying to upload is not a valid backup file

/var/lib/psa/dumps/backup_info_XXX/status_WRONG-FORMAT

Line 2954 error: Element 'template-item', attribute 'name': [facet 'enumeration'] The value 'nginxTransparentMode' is not an element of the set {
.... lots of values ....
}

Can I manually resolve the error in my backup files? I would like to replace nginxTransparentMode to the new key - if existing ...

Thanks in advance

 

[Alexey Baturin]

Hello,

There is some kind of pre-migration issue that stops migration. You need to check info log, which contains a path to pre-migration report, look inside the report and resolve all issues at 'error' level there.
Check this thread http://talk.plesk.com/threads/unable-to-migrate.335670/
There is an usability issue - the pre-migration report should be added to info/debug logs, but it is not added. We will fix that in the next update of Plesk Migrator.

 

[Boindil]

Hello,

There is some kind of pre-migration issue that stops migration. You need to check info log, which contains a path to pre-migration report, look inside the report and resolve all issues at 'error' level there.
Check this thread http://talk.plesk.com/threads/unable-to-migrate.335670/
There is an usability issue - the pre-migration report should be added to info/debug logs, but it is not added. We will fix that in the next update of Plesk Migrator.

Hi,
I've tried again and somehow this time the log file wasn't empty. On previous attempts I already tried that and the file was empty ...

Thanks for the quick answer. I hope the migration succeeds now ...

Anyways these errors should be presented in a better way. Regarding the command line utility I tried replacing the strings in all files (via perl) and then I was able to access the dump via GUI. Sadly Plesk tells me, that my password is not correct - it is... Are the info_*.xml files in some way included in a checksum?

Additionally the nginx errors should be ignored by destination plesk server! I'm trying to transfer from CentOS 6.7 with Plesk 12.5 to CentOS 7 with Plesk 12.5 but somehow these options seem to produce errors on the destination server. The "nginxTransparentMode" option is indeed missing in the subscription settings, but "nginxServeStatic" is configurable - and still produces errors. Please fix these errors.

Greetings,

Adrian

 

[Alexey Baturin]

Anyways these errors should be presented in a better way.

Definitely agree, it is not convenient to look into log files to see what's happened.

Regarding the command line utility I tried replacing the strings in all files (via perl) and then I was able to access the dump via GUI. Sadly Plesk tells me, that my password is not correct - it is... Are the info_*.xml files in some way included in a checksum?

Have you replaced the string before you imported the dump (in archive file), or after on filesystem (at DUMP_D directory, usually /var/lib/psa/dumps)? Plesk adds signature to each backup (http://kb.odin.com/en/116395), but I'm not sure how this could be related to incorrect password. It would be very helpful if you provide step-by-step details, like you did in the first message.

Anyway, if you are trying to restore backup for migration purposes, I would recommend you to use Plesk Migrator instead.

 

[Boindil]

Definitely agree, it is not convenient to look into log files to see what's happened.


Have you replaced the string before you imported the dump (in archive file), or after on filesystem (at DUMP_D directory, usually /var/lib/psa/dumps)? Plesk adds signature to each backup (http://kb.odin.com/en/116395), but I'm not sure how this could be related to incorrect password. It would be very helpful if you provide step-by-step details, like you did in the first message.

Anyway, if you are trying to restore backup for migration purposes, I would recommend you to use Plesk Migrator instead.

Yup, I am restoring the destination server right now to start clean again (tried a lot with the backup/migration). Hopefully I am now able to resolve the issues and use the migrator.

I imported the backup through command line via

/usr/local/psa/bin/pleskrestore

using the provided password.

After that used the following command in /var/lib/psa/dumps:

find . -type f -name '*.xml' -exec \
perl -0777 -pi -e 's#\n  <setting>\n    <name>nginxTransparentMode</name>\n    <value encoding="base64">ZmFsc2U=</value>\n  </setting>\n#\n#igs' {} +

Greetings

Adrian

 

[Boindil]

The Migration manager is having the following problems:

[INFO] Create hosting addon plan '+++ Addon_1' owned by admin
[ERROR] Failed to create hosting addon plan '+++ Addon_1' of admin: Plesk error [1019]: PHP handler id cannot be specified because PHP support is not enabled
[ERROR] Failed to create hosting plans
Cause: Failed to create the following plans on target panel:
- hosting addon plan '+++ Addon_1' owned by admin

The addons do not have any PHP settings. This issue was resolved by creting all addons manually.

Now the manager does create all customers successfully but then fails:

[ERROR] Failed to perform an action on subscription 'domain': Failed to install APS package 'WordPress-4.0-41'. Target panel will not consider this application as installed. However all application-related content should be restored correctly. Exception message: APS package was not found on the source Plesk server
[INFO] Import backup dumps to target panel's repository
[INFO] Restore Apache restart interval
[INFO] Force Apache restart
[INFO] Restore security policy on Windows nodes
[ERROR] Failed to import backup dumps to target panel's repository
Cause: Failed to import backup XML: errcode must be '0' (no errors) or '116' (backup sign error).
Output of pmmcli utility:
<?xml version="1.0" encoding="UTF-8"?>
<response>
    <errcode>152</errcode>
    <errmsg>Unable to import file as dump: The file you are trying to upload is not a valid backup file</errmsg>
</response>

That is a critical error, migration was stopped.

Why the heck does plesk think that the file is not valid? I'm using the migration manager! It HAS to be valid ...