"Failed to upload the migration agent"

[tempest]

Can anyone please help, this is very urgent.

I am trying to move several domains/sites from my client's old server to a new one. Both have Plesk Reloaded (7.5 i think, but not sure how to check). Anyway, I am trying to follow the documentation and trying to use Migration Manager to move the domains over.

However, when I try to upload the migration agent from the new server to the old one, using root as the account, it says:

Failed to upload the migration agent
Cannot transfer agent to the remote host
Lost connection with the remote host

It's giving me absolutely no more clues as to why this is happening. I don't even know how the migration manager is attempting to connect to the old server - is it via ftp or ssh or other? Perhaps it's something to do with ssh access or something?

Looking forward to some help. :/
Thanks in advance,
tempest.

 

[tempest]

well.. i have managed to temporarily allow ssh access on the old host, and now the Migration Manager on the new host seems to be connecting, but is forever refreshing the "Uploading Migration Agent" page..... not sure how long its supposed to take, but ive been here 20 minutes now...

i really hope i can get this to work...

 

[dm__@]

Originally posted by tempest
now the Migration Manager on the new host seems to be connecting, but is forever refreshing the "Uploading Migration Agent" page..... not sure how long its supposed to take, but ive been here 20 minutes now...

AFAIR, it was a very short operation to upload the agent. Does the PleskX.pl exist in /root/migration/<some-number-here> or /migration/<some-number-here> directory on the old host? Try to start it with ./PleskX -s.

BTW, there is log of migration in /usr/local/psa/PMM/var/migration.log on the new host, it contains detailed information about the process. Unfortunately, log file looks really 'debuggish'.

 

[tempest]

I have looked in /root/ and there is no sign of any migration dir.

This is my migration.log, but as you said, its very "debuggish"... The only line I can see which might be a clue is where it says 'data[No suitable authentication method found'. What could this mean? I have temporarily enabled ssh for root, and can login using putty, so that can't be it can it? Or is it that both servers have to have ssh enabled?

I really need help.

[12:05:08 19.05.2005 | pid 27777]eployer | Started (pid 27777, uid 0, euid 0) | '/usr/local/psa/admin/bin/deployer' '--get-fs'
[12:05:08 19.05.2005 | pid 27777]eployer | Finished (pid 27777)
[12:05:55 19.05.2005 | pid 27829]eployer | Started (pid 27829, uid 0, euid 0) | '/usr/local/psa/admin/bin/deployer' '--get-fs'
[12:05:55 19.05.2005 | pid 27829]eployer | Finished (pid 27829)
[12:05:55 19.05.2005 | pid 27846]:Launchpad | Started (pid 27846, uid 0, euid 0) | '/usr/local/psa/admin/bin/launchpad' '--send-scout' '--daemon' '--ip=212.101.266.18' '--login=root'
[12:05:55 19.05.2005 | pid 27848]:Launchpad | (pid 27846) | Has gone to the background (pid 27848)
[12:05:55 19.05.2005 | pid 27848]:1: (in, out, err) Executing command 'perl' on the remote host.
[12:05:55 19.05.2005 | pid 27848]:1: perl
[12:05:55 19.05.2005 | pid 27849]:Executing '/usr/local/psa/admin/sbin/ssh-lite' '-x' '-p' '22' '-oForwardAgent no' '-oClearAllForwardings yes' '-noauth' '-lroot' '212.101.266.18' 'perl'
[12:05:55 19.05.2005 | pid 27848]:1 : 'ssh-lite' process with pid 27849 created
[12:05:55 19.05.2005 | pid 27848]:1: Reading response from remote host
[12:05:55 19.05.2005 | pid 27855]eployer | Started (pid 27855, uid 0, euid 0) | '/usr/local/psa/admin/bin/deployer' '--check-process' '--pid=27848'
[12:05:55 19.05.2005 | pid 27855]eployer | Finished (pid 27855)
[12:05:55 19.05.2005 | pid 27848]:1; readResponse|started |****************************************
[12:05:55 19.05.2005 | pid 27848]:1; readResponse|poll |waiting 2 handles beginning from 0
[12:05:55 19.05.2005 | pid 27848]:1; readResponse|poll |2 handles have data
[12:05:55 19.05.2005 | pid 27848]:1readResponse| handle 0|(13) POLLHUP detected. Closing handle.
[12:05:55 19.05.2005 | pid 27848]:1; readResponse|handle 1|(15) read 41 bytes
[12:05:55 19.05.2005 | pid 27848]:1; readResponse|handle 1|(15) data[No suitable authentication method found^M
]
[12:05:55 19.05.2005 | pid 27848]:1readResponse| handle 1|(15) POLLHUP detected. Closing handle.
[12:05:55 19.05.2005 | pid 27848]:1; readResponse|finish |****************************************
[12:05:55 19.05.2005 | pid 27848]:1: Closing connection
[12:05:55 19.05.2005 | pid 27848]:1: destroying object.
[12:05:55 19.05.2005 | pid 27848]:Launchpad | Finished (pid 27848)
[12:11:21 19.05.2005 | pid 27973]:Launchpad | Started (pid 27973, uid 0, euid 0) | '/usr/local/psa/admin/bin/launchpad' '--send-scout' '--daemon' '--ip=212.101.266.18' '--login=root'
[12:11:21 19.05.2005 | pid 27975]:Launchpad | (pid 27973) | Has gone to the background (pid 27975)
[12:11:21 19.05.2005 | pid 27975]:1: (in, out, err) Executing command 'perl' on the remote host.
[12:11:21 19.05.2005 | pid 27975]:1: perl
[12:11:21 19.05.2005 | pid 27976]:Executing '/usr/local/psa/admin/sbin/ssh-lite' '-x' '-p' '22' '-oForwardAgent no' '-oClearAllForwardings yes' '-noauth' '-lroot' '212.101.266.18' 'perl'
[12:11:21 19.05.2005 | pid 27975]:1 : 'ssh-lite' process with pid 27976 created
[12:11:21 19.05.2005 | pid 27975]:1: Reading response from remote host
[12:11:21 19.05.2005 | pid 27988]eployer | Started (pid 27988, uid 0, euid 0) | '/usr/local/psa/admin/bin/deployer' '--check-process' '--pid=27975'
[12:11:21 19.05.2005 | pid 27988]eployer | Finished (pid 27988)
[12:11:31 19.05.2005 | pid 28031]eployer | Started (pid 28031, uid 0, euid 0) | '/usr/local/psa/admin/bin/deployer' '--check-process' '--pid=27975'
[12:11:31 19.05.2005 | pid 28031]eployer | Finished (pid 28031)
[12:11:42 19.05.2005 | pid 28074]eployer | Started (pid 28074, uid 0, euid 0) | '/usr/local/psa/admin/bin/deployer' '--check-process' '--pid=27975'
[12:11:42 19.05.2005 | pid 28074]eployer | Finished (pid 28074)
[12:11:52 19.05.2005 | pid 28117]eployer | Started (pid 28117, uid 0, euid 0) | '/usr/local/psa/admin/bin/deployer' '--check-process' '--pid=27975'
[12:11:52 19.05.2005 | pid 28117]eployer | Finished (pid 28117)
[12:12:03 19.05.2005 | pid 28160]eployer | Started (pid 28160, uid 0, euid 0) | '/usr/local/psa/admin/bin/deployer' '--check-process' '--pid=27975'
[12:12:03 19.05.2005 | pid 28160]eployer | Finished (pid 28160)
[12:12:13 19.05.2005 | pid 28203]eployer | Started (pid 28203, uid 0, euid 0) | '/usr/local/psa/admin/bin/deployer' '--check-process' '--pid=27975'
[12:12:13 19.05.2005 | pid 28203]eployer | Finished (pid 28203)

The last few lines are repeated over and over with different pids, presumable in conjuction with the page refreshes on Plesk.

Can anyone see the problem??

 

[dm__@]

Originally posted by tempest
authentication method found'. What could this mean? I have temporarily enabled ssh for root, and can login using putty, so that can't be it can it? Or is it that both servers have to have ssh enabled?
[/B]

I've seen the 'No suitable ...' message in logs only when I passed the wrong password, so it seems really to be an authentication problem. May be Migration Manager and remote ssh are trying to use different authentication methods (keyboard-interactive, and password, and whatever else SSH protocol allows)? BTW, I've found that /usr/local/psa/admin/sbin/ssh-lite (mentioned in the log above) is like regular 'ssh' (I don't know why does this exist, may be some patches they wanted or just compatibility issues), so you can try to play with this utility to see can it connect to the your old server, and if it can't you may try to contact support team - may be it's a bug in Migration Manager, not the unusual ssh daemon settings.

And, yes, I migrated _to_ the server with carefully firewalled ssh. Everything went fine.

 

[Chris]

Try SSH-ing as root from one server to the other to verify you can log in. Also make sure wget is enabled on the server that migration manager is trying to log in to.

 

[dm__@]

Originally posted by Chris
Also make sure wget is enabled on the server that migration manager is trying to log in to.

Nope, wget is not needed.

 

[rank1st]

Just a short note that I thought I would add after scratching my head for an hour:

Check the "Other" radio button and put in the username root. It worked for me.