koha-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Koha-devel] various questions (mainly around zebra config & stuff)


From: Paul POULAIN
Subject: [Koha-devel] various questions (mainly around zebra config & stuff)
Date: Fri, 04 Jan 2008 11:48:04 +0100
User-agent: Thunderbird 2.0.0.6 (X11/20070914)

~#perl -e 'print "Hello world\n";

Investigating liblime recent or less recent commits I have some questions :
==== 1 ====
- the koha-conf.xml now contains
<listen id="biblioserver" >unix:__ZEBRA_RUN_DIR__/bibliosocket</listen>
so, no more address@hidden:port

iirc, unix sockets are much more faster than tcp.
but :
- does that means we "loose" the z3950 public server in the process ?
- Is it replaced by the SRU/SRW that is just after ? (ie : can we query the SRU/SRW server with PQF queries ?)

==== 2 ====
How to check that the SRU/SRW works fine ? I tried a yaz-client @:9998/Biblios and get :
Connecting...error = System (lower-layer) error: Connection refused

==== 3 ====
The opensearch & unapi things uses SRU/SRW, but the port is hardcoded to 9998, were it can be changed in the install process.
can we consider that as a bug ?

==== 4 ====
misc/zebraqueue_start.pl & misc/update_items.pl
afaik :
- zebraqueue_start was written to update the zebra DB when a biblio is updated/added/deleted - update_items was written to update on a daily basis all items that have been updated (to update issuing status of the item)

now that zebra is quick, a line is added into zebraqueue and the "position status" of each item is updated everytime zebraqueue_start is run.
so do we still need misc/update_items.pl ?

==== 5 ====
It seems that issuing a book results in 2 lines in zebraque. Isn't this a bug ?

==== 6 ====
Now that zebra indexes quickly, do we still need zebraqueue_start ? couldn't we update in real time ? (that wouldn't solve the multiple access problem joshua faced 2 years ago, when updating the DB from 2 processes resulted in a DB crash. But myabe this issue is solved as well now.)

==== 7 ====
couldn't we transform zebraqueue_start to a daemon, and not let is as cronjob ? having to wait for 1mn in the worst case seems very long to most librarians (when they add an authority and have to wait for 1mn to be able to find it in the biblio)
would that daemon-ization be a problem for MS-windows servers ?

--
Paul POULAIN
BibLibre SARL
Expert en Logiciels Libres pour l'info-doc
Tel : 04 91 31 45 19




reply via email to

[Prev in Thread] Current Thread [Next in Thread]