tonyair/silverstripe-fulltextsearch
1
0
Fork 0
mirror of https://github.com/silverstripe/silverstripe-fulltextsearch synced 2024-10-22 14:05:29 +02:00
Code Issues Projects Releases Wiki Activity

NEW Ensure commits are visible to seachers (fixes #274)

Browse Source 
This is required in situations where autoCommit.openSearcher is set to "false",
and you don't want to rely on core restarts to make committed index updates available to new searches.
It's already configured like this in the Common Web Platform environments,
so this change will mainly affect other environments (such as Silverstripe Platform).
To opt out of this new behaviour, set up a custom solrconfig.xml.

More details in https://github.com/silverstripe/silverstripe-fulltextsearch/issues/274
This commit is contained in:
Ingo Schommer 2020-04-07 17:22:29 +12:00
parent 4be5de76a5
commit 8876a47f0d
2 changed files with 193 additions and 180 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
conf/solr/4/extras/solrconfig.xml
View File

@ -379,7 +379,7 @@
-->
<autoSoftCommit>
<maxTime>${solr.autoSoftCommit.maxTime:-1}</maxTime>
<maxTime>${solr.autoSoftCommit.maxTime:60000}</maxTime>
</autoSoftCommit>
<!-- Update Related Event Listeners

17
docs/en/06_troubleshooting.md
View File

@ -1,13 +1,26 @@
# Troubleshooting
## Newly indexed content only shows in searches after a delay
First, check how you're running index operations.
In many cases where the `queuedjobs` module is installed,
saving or publishing a record will create a new index job which needs to complete first.
Solr also distinguishes between adding documents to the indexing,
committing them, and making them available to new searches.
In most cases this happens within a few seconds, but
in sometimes it can take up to a minute due to the
`autoSoftCommit` configuration setting defaults in your `solrconfig.xml`.
To find out more detail, read about
[soft vs. hard commits](https://lucidworks.com/post/understanding-transaction-logs-softcommit-and-commit-in-sorlcloud/).
## Common gotchas
* By default number-letter boundaries are treated as a word boundary. For example, `A1` is two words - `a` and `1` - when Solr parses the search term.
* Special characters and operators are not correctly escaped
* Multi-word synonym issues
* When Dolr indexes are reconfigured and reindexed, their content is trashed and rebuilt
* When Solr indexes are reconfigured and reindexed, their content is trashed and rebuilt
### CWP-specific
## CWP-specific
* `solrconfig.xml` customisations fail silently
* Developers aren’t able to test raw queries or see output via the