This page last updated: 01/14/08 12:01pm
3.1
New features
- "=" is now interchangeable with "::" when using search fields in a search.
- Users can now type values in a form search.
- New commands resurrect , unresurrect, and export are available in the CLI:
- resurrect: makes data available that has previously been archived.
- unresurrect: used to delete directories that have been resurrected.
- export: exports user data, global data, or event data for archiving.
- For help with these commands, access the CLI help pages through the command line: ./splunk help resurrect or ./splunk help unresurrect or ./splunk help export.
Resolved issues from 3.0.x
- SplunkWeb now performs searches faster with smaller result sets.
- SplunkWeb no longer elevates user privileges without warning.
- Whitelisting and blacklisting now work correctly.
- Distributed searches now access events in parallel across all servers.
- SplunkWeb show source now displays correctly.
- Dashboard Saved searches in SplunkWeb now display correctly.
- When selecting fields in a search in SplunkWeb, the selected fields show correctly.
- Splunk no longer suffers a performance loss when extracted and search fields are enabled.
- Unprintable characters no longer cause errors in the CLI.
- Unprintable characters return as "?".
- Original data is not lost. Use the extract command to retrieve the raw data.
- Entries are no longer dropped for LDAP servers running Active Directory.
- Splunk's log volume is now significantly smaller.
- Setting sourcetypes of archived files works properly.
- Fixed small bugs with saved searches and alerts in distributed search deployments.
- Splunkd is more efficient.
- Event type tagging works correctly in SplunkWeb.
- Deleting events while a source is being indexed no longer causes an incorrect event count to be reported.
- Splunk runs correctly on AIX 5.2 and 5.3.
- Splunk has Improved handling of TCP inputs.
- Shutting down Splunk does not cause TCP event loss.
- Splunk functions correctly with the gzip environment variable set.
- CSV files with common headers no longer cause Splunk's tail to miss events.
- Splunk has increased functionality with Internet Explorer 6.
- Fixed cosmetic bugs in SplunkWeb.
- eventtypetag:: typeahead is corrected.
- hosttag:: tyepahead is corrected.
- Runtime field extraction works with host::.
- Improved user role capabilities.
- Enhanced search performance on 64-bit systems.
- Dashboard creation now works correctly in SplunkWeb.
- Selecting "table" from drop-down in the SplunkWeb dashboard does not produce both a chart and a table.
- Alerts created by power users now can perform all actions as expected
- Compressed files now correctly handle sourcetypes manually set in inputs.conf
- Dashboard no longer displays duplicate sets of example searches for every distributed data source.
- Full TCP input queues no longer drops events.
- The search command outputcsv outputs fields in the proper order.
New issues in this release
- In a distributed search cluster, it is recommended that you have every node upgraded to an identical version of Splunk.
- You can never mix 3.1.x and 3.0.x nodes in a distributed search cluster. You must upgrade all 3.0.x nodes to 3.1.x.
- When shutting down Splunk, there is a risk that network inputs can experience event loss.
- Having too many LDAP roles set might cause searching to suffer a performance loss.
- Using time-based search modifiers in the format: modifier::value in a saved search will break links that are sent in alerts via RSS or email.
- Fix this by changing all custom time-based modifiers used in savedsearches.conf to: timebasedmodifier=value or timebasedmodifier="value" format.
- Searches using a literal equal sign (=) will no longer work due to changes in the search syntax. This may cause saved searches to fail.
- Fix this by enclosing the search expression in double quotes. For example, "user=foo" .
Comments
No comments have been submitted.