SRM2 - Bug Fixes and Enhancements for v2.9-1

Release Manager: Giuseppe Lo Presti
Planned Release Date: 22 Feb 2010
Code Freeze Date: 31 Jan 2010
Downtime required: O(mins)
Release Version Number: 2.9-1

Max. Runtime for upgrade scripts O(secs)

Bug Bug Title Requires Schema Change Transparent Upgrade Hotfix Candidate Status [Planned / Developed / Testing / Confirmed]
#45082 RFE: change interface for blocking stager calls to use a common callback YES NO NO CONFIRMED
#60036 RFE: on prepareToGet without space token, use the ability to find the service class as provided by stage_filequery NO NO NO CONFIRMED
#60503 RFE: expose space token on srmLs when file is ONLINE NO YES NO CONFIRMED
#61250 SRM does not respect desiredTotalRequestTime on PToGet and BoL NO YES NO CONFIRMED
#61905 setPermission not working NO YES NO CONFIRMED / BP
#62271 SRM should perform all queries to the NameServer on behalf of the mapped user NO YES NO CONFIRMED / BP
#62972 srmRm should delegate permissions checks to Cns_unlink NO YES NO CONFIRMED / BP
N/A Support for SLC5 NO YES NO CONFIRMED

Deployment notes

  • This release has a known regression bug on the srmCopy functionality, which leads the backend daemon to core dump on SLC4 because of a memory corruption in the globus libraries. Note however that the SLC5 version is immune to this problem, hence should the copy functionality be needed, an upgrade to SLC5 is required. The recommended workaround in SLC4 is to disable the copy threads in srm2.conf.
  • Due to a number of fixes in the Castor backend (cf. the release notes), the minimum required release of the underlying Castor backend is either 2.1.8-18 or 2.1.9-4. Any earlier Castor release (including the 2.1.7 series) is not supported with this release of SRM. Note that the service won't refuse to run but failures may not get properly propagated in some circumstances.
  • The fix for bug #45082 (RFE: change interface for blocking stager calls to use a common callback) requires the RH/SRMHostsList configuration option turned on in the castor.conf of the underlying Castor instance (refer to the release notes for details). This parameter has been provided since Castor 2.1.8-7. In case this is missing, the SRM service will systematically fail all transfer requests.
  • The default stager client timeout has been shortened to 20 seconds by default. This applies to all fast (i.e. non-scheduled) requests. The scheduled requests are now handled asynchronously without any timeout.

-- GiuseppeLoPresti - 11-Jan-2010

Edit | Attach | Watch | Print version | History: r29 < r28 < r27 < r26 < r25 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r29 - 2010-02-18 - GiuseppeLoPresti
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    DataManagement All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2023 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback