Wide character in print at /usr/bin/pt-archiver line 6700.

[le_den]

@carlos.salguero;52217

Hello,

Thanks for you patience and your willing to help.
May I ask you to do one more test please?

Add 2 binmode lines as follows: (lines 6649 & 6650)

6646 my $need_hdr = $o->get('header') && !-f $archive_file;
6647 $archive_fh = IO::File->new($archive_file, ">>$charset")
6648 or die "Cannot open $charset $archive_file: $OS_ERROR\n";
6649 binmode STDOUT, ":utf8";
6650 binmode $archive_fh, ":utf8";
6651 $archive_fh->autoflush(1) unless $o->get('buffer');

Regards

Hi Carlos,

That helped, the tool is running now without stopping due to reported error. I’ll report back in case it reoccurs.

Thanks for your help, much appreciate it!

[carlos.salguero]

Hi,

Please let me know if it works and if it doesn’t work. I want to officially add the fix.

Thanks for your help.

[le_den]

Hi,

That works.

Once I added two lines that you asked to add the tool is running since then without any issue, It’s currently archiving DB that have reported characters.

I want to officially add the fix.

Great, thanks.

[carlos.salguero]

Hello,

Just a follow up. I’ve created [url][PT-1611] pt-archiver fails with UTF-8 chars - Percona JIRA
It will be released in next PT (3.0.12)

Regards

[Tobias_Brox1]

I got the same now with pt-archiver 3.5.0, after replacing the source database from a percona database to a mariadb database. The default character set was also changed. Perhaps it’s related to the mysql-specific charset utf8mb4.

I managed to work around the problem with this hack:

--- /tmp/pt-archiver-orig	2023-01-04 14:02:54.597644218 +0100
+++ /tmp/pt-archiver	2023-01-04 14:02:42.031017165 +0100
@@ -6642,7 +6642,7 @@
       return 0;
    }

-   my $charset  = $got_charset || '';
+   my $charset  = 'utf8';
    if ($charset eq 'utf8') {
       $charset = ":$charset";
    }

[Tobias_Brox1]

A more proper fix would probably be to add a --charset-option