LP# 1330784: Release notes for sitemap builder

More like documentation than release notes, but more is probably better
than less.

Signed-off-by: Dan Scott <dscott@laurentian.ca>
Signed-off-by: Ben Shum <bshum@biblio.org>

diff --git a/docs/RELEASE_NOTES_NEXT/OPAC/sitemap_builder.txt b/docs/RELEASE_NOTES_NEXT/OPAC/sitemap_builder.txt
new file mode 100644 (file)
index 0000000..527a5ed
--- /dev/null
+++ b/docs/RELEASE_NOTES_NEXT/OPAC/sitemap_builder.txt
@@ -0,0 +1,51 @@
+Sitemap generator
+^^^^^^^^^^^^^^^^^
+A http://www.sitemaps.org[sitemap] directs search engines to the pages of
+interest in a web site so that the search engines can intelligently crawl
+your site. In the case of Evergreen, the primary pages of interest are the
+bibliographic record detail pages.
+
+The sitemap generator script creates sitemaps that adhere to the
+http://sitemaps.org specification, including:
+
+* limiting the number of URLs per sitemap file to no more than 50,000 URLs;
+* providing the date that the bibliographic record was last edited, so
+  that once a search engine has crawled all of your sites' record detail pages,
+  it only has to reindex those pages that are new or have changed since the last
+  crawl;
+* generating a sitemap index file that points to each of the sitemap files.
+
+Running the sitemap generator
++++++++++++++++++++++++++++++
+The `sitemap_generator` script must be invoked with the following argument:
+
+* `--lib-hostname`: specifies the hostname for the catalog (for example,
+  `--lib-hostname https://catalog.example.com`); all URLs will be generated
+  appended to this hostname
+
+Therefore, the following arguments are useful for generating multiple sitemaps
+per Evergreen instance:
+
+* `--lib-shortname`: limit the list of record URLs to those which have copies
+  owned by the designated library or any of its children;
+* `--prefix`: provides a prefix for the sitemap index file names
+
+Other options enable you to override the OpenSRF configuration file and the
+database connection credentials, but the default settings are generally fine.
+
+Note that on very large Evergreen instances, sitemaps can consume hundreds of
+megabytes of disk space, so ensure that your Evergreen instance has enough room
+before running the script.
+
+Scheduling
+++++++++++
+To enable search engines to maintain a fresh index of your bibliographic
+records, you may want to include the script in your cron jobs on a nightly or
+weekly basis.
+
+Sitemap files are generated in the same directory from which the script is
+invoked, so a cron entry will look something like:
+
+------------------------------------------------------------------------
+12 2 * * * cd /openils/var/web && /openils/bin/sitemap_generator
+------------------------------------------------------------------------