Although more testing is appropriate I believe statistics it is now.
Thank you for the link posted by Hilton Gibson on rebuilding the
statistics from scratch.
At our site (Sydney University) when we upgrade or create a new instance
of DSpace for some new purpose, we usually make a clone of a working
machine. We then give it a new IP and name, enter the name and IP into
the DNS alter the hosts and DSpace config files to reflect the new name
then do the upgrade (or whatever).
Running the unix command "strings: on the stats database files in
<dspaceRoot>/solr/statistics/data/index/data
there are references to the old machine name and this suggested to me
this was a possible source of the problems and we needed to rebuild
statistics within the new machine identity.
The link from Hilton is a little out dated the new procedure should be
1. Backup <dspaceRoot>/solr/statistics/data
2. Delete <dspaceRoot>/solr/statistics/data
4. Reboot your machine
5. Convert old log files
cd <dspaceRoot>/log
ITEM=`ls dspace.log.*`
for i in $ITEM ; do
<dspaceRoot>/bin/dspace stats-log-converter -i $i -o $i.solr
done
6. Import converted log files
cd <dspaceRoot>/log
ITEM=`ls *.solr`
for i in $ITEM ; do
<dspaceRoot>/bin/dspace stats-log-importer -i $i
done
The statistics counts on any particular item appear correct and are updating
counts of test accesses.
--
Keir Vaughan-Taylor
ERA DSpace Administrator
Rm:123 Fisher Library University of Sydney
[email protected]
ph:+61 2 9351 7408
mb:+61 434 606 885
------------------------------------------------------------------------------
_______________________________________________
DSpace-tech mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dspace-tech
List Etiquette: https://wiki.duraspace.org/display/DSPACE/Mailing+List+Etiquette
