X-Git-Url: http://jsfdemo.indexdata.com/?a=blobdiff_plain;f=archive%2Ftest-server%2FREADME;h=15e4f736239b493cec7e7c1d9474d45f63f4275e;hb=bffbc733bde1e0399476faa10724af389cab586c;hp=d1bcab45f7e361d6e470fde9278f430e8be00614;hpb=f489c5ea8d36090d4f91714208bde4b4214d7d9b;p=ZOOM-Perl-moved-to-github.git

diff --git a/archive/test-server/README b/archive/test-server/README
index d1bcab4..15e4f73 100644
--- a/archive/test-server/README
+++ b/archive/test-server/README
@@ -1,5 +1,3 @@
-$Id: README,v 1.2 2008-09-26 12:23:26 mike Exp $
-
 These files are used to run the server that the ZOOM-Perl test-suite's
 "packages" scripts run against.
 
@@ -10,3 +8,25 @@ test-server.init -- script to run the test server from /etc/init.d
 test-server.cfg -- Zebra configuration file for the test server.
 test-server.pw -- password file for test server's Zebra configuration.
 
+--
+
+Over time, the Zebra database where the ZOOM-Perl package read-write
+tests take place becomes bloated and unclean: when I checked on 15
+October 2010, there were over 15000 "sort<NUMBER>-0.mf" files!  As a
+result, the test-suite becomes slower, to the point where
+26-packages.t takes two and half minutes to run.
+
+When this happens, reset the database to a clean state:
+
+$ ssh mike@z3950.indexdata.com
+$ cd ZOOM-Perl/archive/test-server
+$ sudo /etc/init.d/zoom-perl-server stop
+$ sudo zebraidx-2.0 -c test-server.cfg init
+$ sudo zebraidx-2.0 -c test-server.cfg update ../../../idzebra/examples/gils/records
+	(This may take a few minutes!)
+$ # If desired, test with: zebrasrv-2.0 -c test-server.cfg
+$ sudo /etc/init.d/zoom-perl-server start
+$ exit
+
+Once this has been done, the time to run 26-packages.t should drop
+back down to four or five seconds!