Hi,

Thanks to everyone who have been helping me to sort this out.  Its pretty
much there now.

I only have one other issue that crops up occasionally and completely
randomly.  

When I run the sync script I normally get the following output:

repligard-Message: Reading config file
/var/lib/aegir/etc/repligard_hourly_dump_staging.conf
repligard-Message: Exporting to /tmp/hourly_export_200405101101.xml.gz
Decompressing XML Archive
Removing unapproved elements
Removing expired elements
Converting xml file to UTF-8
repligard-Message: Reading config file
/var/lib/aegir/etc/repligard_hourly_import_live.conf
repligard-Message: Importing /tmp/hourly_export_200405101101_conv.xml

On occasion, especially when styles are concerned or the data is very large,
I get the following instead.  Nothing is transferred to the live server and
anything approved on the staging server does not get transferred if I rerun
the command (which I would expect given what the export script does)

repligard-Message: Reading config file
/var/lib/aegir/etc/repligard_hourly_dump_staging.conf
repligard-Message: Exporting to /tmp/hourly_export_200405101001.xml.gz
Decompressing XML Archive
Removing unapproved elements
Removing expired elements
Converting xml file to UTF-8
iconv: illegal input sequence at position 6390
repligard-Message: Reading config file
/var/lib/aegir/etc/repligard_hourly_import_live.conf
repligard-Message: Importing /tmp/hourly_export_200405101001_conv.xml

REPLIGARD/READ-CRITICAL **: unclosed CDATA section in line 128
<------ This is the line that crops up at random

Can anyone shed any light on this?

Cheers

Mike.

-----Original Message-----
From: Eero af Heurlin [mailto:[EMAIL PROTECTED]
Sent: 26 April 2004 15:02
To: [EMAIL PROTECTED]
Subject: Re: [midgard-user] Staging / Live Server Synchronisation
Problem


Michael Ross (PCT North West) wrote:
>  Is there not a faster and less
> processor/memory intensive way of doing this as this makes the server with
> all our live sites  so slow to respond it is almost unusable - the perl
> script hogs 98% of the processor and all the available ram.
> 

The old Nadmin scripts were faster but they also were 1) horrid hack 
(manipulate the database timestamps directly to "force" or "prevent" 
replication) 2) very unreliable.

You can use the nice-command to run the script at lower priority (nice 
-n15 remove_unapproved.pl dump.xml), however this does not prevent it 
from hogging the ram it needs (which will be in the range of the size of 
the input file before it's finished), increasing the available swap (or 
preferably ram) is the only solution to memory running out alltogether.

/Rambo

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to