[SOLVED] UTF8 error in cluster sync

[tmwl]

Hello,

I set up a mail cluster and I have a little problem with one of the nodes. The master (mx1) has no sync errors in the log but the node mx2 seems to have problems:

 starting cluster syncronization
Mar  4 10:00:33 mx2 pmgmirror[1333]: detected rule database changes - starting sync from 'xxx.xxx.xx.xx'
Mar  4 10:00:33 mx2 pmgmirror[1333]: pg_putcopyend cannot be called until a COPY is issued at /usr/share/perl5/PMG/DBTools.pm line 1015.
Mar  4 10:00:33 mx2 pmgmirror[1333]: database sync 'mx1' failed - DBD::Pg::db pg_putcopyend failed: ERROR:  invalid byte sequence for encoding "UTF8": 0xfc#012CONTEXT:  COPY objectgroup, line 32 at /usr/share/perl5/PMG/DBTools.pm line 1012.
Mar  4 10:00:33 mx2 pmgmirror[1333]: cluster syncronization finished  (1 errors, 0.34 seconds (files 0.00, database 0.18, config 0.15))

Maybe you have a solution for this problem.

 

[mira]

This might be related to: https://bugzilla.proxmox.com/show_bug.cgi?id=1978 that was fixed in 5.1-3
Try recreating the slabve db with 'pmgdb delete' followed by 'pmgdb init' or remove the slave and add it again after upgrading to the newest version.

 

[tmwl]

Thanks for the fast answer. I updated my version (This is a test cluster without subscription so I forgot to add

deb http://download.proxmox.com/debian/pmg stretch pmg-no-subscription

to the apt repository list), deleted the node with

pmgcm delete 3

(pmgdb delete throws an error), added the node back and the error is fixed.

 

[Rafael Solano]

Thanks for this answer Mira, I was also facing the same issue regarding this.