What happened to RIPE Database full text search?

Hi Cleve,

 "Cleveland K. Mickles" <cmickles at mci.net> writes:
 * 
 * We had a problem with customers using RIPE last night.  Was 
 * there a change this past weekend?

A script was missing and that caused the files in the ftp site to be empty.

 * 
 * We copied over the RIPE database and it is taking longer
 * than normal to process.  Is indexing information
 * usually passed along with the FTP copy of the RIPE 
 * database?

No, indexing information is not contained on the ftp files. They contain only 
the data itself. You have to index them after ftp-ing them to your machine.
Indexes are dependent on your PERL setup (eg. if you index the database with 
perl5.003, a whoisd using perl5.004 won't be able to use them), so there is no 
point in distributing the indexes.

The files on the ftp site were restored at 12:00 (our time). What you got 
depends on the time you got it.

Anyway it should not be taking longer than normal because you either got an 
empty file or a correct one (unless you happened to ftp the file at the same 
time we were putting it in our ftp site and you have a truncated file. Then 
the indexing program can choke. Have you checked file length against the one 
in our ftp site?)

Now almost everything is back to normal, including the full search text search.
The missing bit is the full version of the database file with the attributes 
translated to the long version. This file takes so long to generate that we 
calculated by the time it was finished, the process to generate today's one 
would be starting so we preferred to get the full search text indexing done 
first.

Apologies for any incovenience.

Regards,
Joao Damas
RIPE NCC

 * 
 * Regards,
 * 
 *    Cleve...
 *