Thursday, August 26, 2010

OSG 1.2.13 Release Announcement

OSG Operations and Integration are pleased to announce the release of OSG version 1.2.13.

The following components are affected:

* Gratia accounting on CE installations
* Xrootd installations
* OSG client and wn-client installations
* Systems using gratia rpms

This release has updates to three components. Xrootd, Gratia, and the Fermi SRM clients have been updated. The updates are fairly minor and do not necessitate an immediate update unless you are encountering an issue that is corrected in these updates.

Notable fixes include correct a syntax issue when specifying the redirector on a data server. Gratia has been updated to correct several bugs and has added accounting probes for Xrootd. Fermi SRM clients have been updated as well.

The release notes for the VDT 2.0.0p20 release underlying this release can be found here: http://vdt.cs.wisc.edu/releases/2.0.0/release-p20.html

Update instructions can be found on the OSG twiki under the OSG 1.2.12: https://twiki.grid.iu.edu/bin/view/ReleaseDocumentation/OSG12UpdateInstructions

Tuesday, August 24, 2010

**UPDATE** - GOC Service Update - Tuesday, August 24th at 14:00 UTC

Earlier today, it was announced that the GOC would briefly bring the TWiki and OIM down today to perform maintenance. Unfortunately, that outage period has been pushed back. We anticipate these outages to be brief and to occur between 14:00 and 17:00 EDT.

We apologize for any inconvenience this may cause.

GOC Service Update - Tuesday, August 24th at 14:00 UTC

The GOC would like to clarify the service announcement for this week
by explicitly announcing the services to be effected by todays
release include TWiki and OIM. We anticipate these outages to be brief
and occur between 9:00 and noon EST.

Monday, August 23, 2010

SAM anomalies reported - **UPDATE** - GOC Ticket # 9120

The WLCG has confirmed with the GOC that Service Availability Monitoring messages are now arriving properly. We are currently re-sending the missed records after receiving the confirmation that the mechanism is fixed and will continue to track this issue to ensure there is no recurrence.

Please see ticket 9120 at:
https://ticket.grid.iu.edu/goc/viewer?id=9120

SAM anomalies reported - GOC Ticket # 9120

This morning, Service Availability Monitoring problems were reported in which sites were showing in unknown status. The GOC is currently giving this issue top priority. Please be advised of this issue and expect another notification when we can confirm with our WLCG collaborators that this issue is fully resolved and we will additionally re-send any data that was not properly received during this time.

We apologize greatly for the inconvenience and thank you for your patience.

Please see ticket 9120 at:
https://ticket.grid.iu.edu/goc/viewer?id=9120

Tuesday, August 17, 2010

GOC Service Update - Tuesday, August 24th at 14:00 UTC

The GOC will upgrade the following services beginning at Tuesday, August 24th, 2010 at 14:00 UTC. The GOC reserves four hours (14:00 - 18:00 UTC) in the unlikely event that unexpected problems are encountered.
GOC Ticket Synchronizer 1.11

* Updated MySQL Connector/J, and fixed load balancing issue with data1/2.
* Updated BNL production RT's default queue name to StorageManagement as requested by Jason
* Added GGUS > FP conversion for GGUS "Ticket Type" field, and FP > BNL converstion.
* (Patched) Fixed the build file issue for missing lib directory.
* (Patched) Added "loadbalance:" token for JDBC URL which was somehow missing.
* Added more logs during sync table connection.

MyOSG 1.25 (https://myosg.grid.iu.edu)

ITB version is now available for testing at https://myosg-itb.grid.iu.edu; we encourage users to test this service before the production release.

* Applied similar update I made for sc and vo page (for better error message when no resource is selected)
* Fixed the subnav style issue.
* Updated the error message displayed when no entity is selected or all entities are filtered out.
* Removed link for Google Wave widget subscription [MYOSG-84]
* Fixed the "expired status" issue which was displayed for resources with OK status. [MYOSG-78]
* Adjusted fontsize and paddings for status map bubble.
* (Patched) Hidden Misc. / OSG User page for non OIM registered users
* (Patched) Changed the VO summary item header so that it reflects the sort type
* Other minor Bug Fixes and cosmetic changes.

GOC Ticket 1.25 (https://ticket.grid.iu.edu)

ITB version is now available for testing at https://ticket-itb.grid.iu.edu; we encourage users to test this service before the production release.

* Navigator: Fixed the issue where column selector sometimes doesn't work.
* Navigator: Hidden security ticket for non OIM registered users.
* Admin: Added assignee override features and implemented GUI. [GOCTICKET-42]
* Other minor updates.


Redhat Software Updates

We will be installing Redhat software updates to all production services except for BDII. This will require reboots in most cases; however, in those cases where we have redundancy via DNS round-robin, we will be making use of this to reduce or eliminate service interruption. Only the updates that have already been installed on the ITB hosts will be installed on the production hosts. The production BDII servers will not be updated in this release.

We will be rebooting is1.grid.iu.edu, which has been responding sluggishly since an attempt to install a software update, which was rolled back. This is one of the two BDII servers, and we will be utilizing the DNS round robin to shift traffic to the other server, which is not being rebooted at this time, so we expect no interruption in service.

Power outage affecting VDT services on Wednesday, August 18

From: Alain Roy
To: OSG GOC
Subject: Power outage affecting VDT services on Wednesday, August 18
Date-Sent: Tuesday, August 17, 2010 3:08 PM -0500

Hi everyone,

There will be a power outage affecting VDT services on Wednesday, August 18 due to upgrades of utilities in our building.

We expect it to begin around 4:00pm Central US time. It will affect:

* The VDT web site at vdt.cs.wisc.edu
This should return in a few hours.

* The VDT software caches (also hosted at vdt.cs.wisc.edu)
This should return in a few hours.

* The VDT ticket system (vdt-support@opensciencegrid.org and crt.cs.wisc.edu)
This may be down until Thursday morning.

We apologize for the downtime and hope it doesn't cause any serious inconveniences for you.

-alain
-----------------------------------------------------------------
Alain Roy
Open Science Grid Software Coordinator roy@cs.wisc.edu
http://opensciencegrid.org http://vdt.cs.wisc.edu

Power outage affecting VDT services on Wednesday, August 18

From: Alain Roy
To: OSG GOC
Subject: Power outage affecting VDT services on Wednesday, August 18
Date-Sent: Tuesday, August 17, 2010 3:08 PM -0500

Hi everyone,

There will be a power outage affecting VDT services on Wednesday, August 18 due to upgrades of utilities in our building.

We expect it to begin around 4:00pm Central US time. It will affect:

* The VDT web site at vdt.cs.wisc.edu
This should return in a few hours.

* The VDT software caches (also hosted at vdt.cs.wisc.edu)
This should return in a few hours.

* The VDT ticket system (vdt-support@opensciencegrid.org and crt.cs.wisc.edu)
This may be down until Thursday morning.

We apologize for the downtime and hope it doesn't cause any serious inconveniences for you.

-alain
-----------------------------------------------------------------
Alain Roy
Open Science Grid Software Coordinator roy@cs.wisc.edu
http://opensciencegrid.org http://vdt.cs.wisc.edu

Tuesday, August 10, 2010

OSG 1.2.12 Release Announcement

OSG Operations and Integration are pleased to announce the release of OSG version 1.2.12.

This update affects all OSG CE installations and SE installations that would like to report to BDII without having a CE.

This release updates GIP to allow GIP to report SE information without having to rely on a CE to report information to the OSG BDII servers. There have also been several bug fixes to the GIP code.

The release notes for the VDT 2.0.0p19 release underlying this release can be found here: http://vdt.cs.wisc.edu/releases/2.0.0/release-p19.html

Update instructions can be found on the OSG twiki under the OSG 1.2.12: https://twiki.grid.iu.edu/bin/view/ReleaseDocumentation/OSG12UpdateInstructions

Wednesday, August 4, 2010

GOC Service Update - Tuesday, August 10th at 14:00 UTC

The GOC will upgrade the following services beginning at Tuesday, August 10th, 2010 at 14:00 UTC. The GOC reserves four hours (14:00 - 18:00 UTC) in the unlikely event that unexpected problems are encountered.

OIM 2.23 (https://oim.grid.iu.edu)

ITB version is now available for testing at https://oim-itb.grid.iu.edu; we encourage users to test this service before the production release.

Release Notes:
Increased the maximum number of characters allowed for downtime editor's summary field, and added a more user friendly validation.
Added check for CLIENT_VERIFY_SSL flag for improved security
Minor cosmetic changes & improved logging.

MyOSG 1.24 (https://myosg.grid.iu.edu)

ITB version is now available for testing at https://myosg-itb.grid.iu.edu; we encourage users to test this service before the production release.

Release Notes:

Added Resource Group / BDII information XML schema which was missing.
Added a new page under Misc. callled "OIM Users" which let user search / show OSG user profiles & contact information [MYOSG-42]
Set correct file name for KML output for Status Map page (for Google Earth)
Added "Promo View" for Status Map which masks unknown site as Greendots

GOC Ticket 1.24 (https://ticket.grid.iu.edu)

ITB version is now available for testing at https://ticket-itb.grid.iu.edu; we encourage users to test this service before the production release.

Release Notes:

Viewer / Added a filter for guest user to hide email address on update "by" entries. [GOCTICKET-64]
Viewer / Fixed the layout issue for guest user
Submitter / Twiki / Added a new field to ask what type of TWiki issue user is having [GOCTICKET-68]
Navigator / Adjusted datatable css again.. and fixed the word-wrapping issue on Chrome for table headerNavigator / Added column specific filters.Added REST / "List Open Tickets" interface for programmatic query of GOC tickets
Home / Added ticket expectation note (GOCTICKET-44)
Patched a warning caused by incorrect query for quick description when null value is selected.
SOAP API / Fixed the issue where closed date information is not correctly set.
Cosmetic updates & improved logging.
Various bug fixesGOC Ticket Synchronizer 1.10Release Notes:Updated MySQL Connector/J, and fixed load balancing issue with data1/2.
Fixed the issue where notification suppression for assignee was disabled causing redundant email notifications sent on some circumstances.Stopping old GGUS ticket exchange script (no longer used by any tickets)

OSG TWiki (https://twiki.grid.iu.edu)

ITB version is now available for testing at https://twiki-test.grid.iu.edu; we encourage users to test this service before the production release.

Release Notes:
Added new style for CEMon/BDII (http://is.grid.iu.edu)

There has been a security update to OpenLDAP; Redhat has backported a patch to OpenLDAP while maintaining the same major version, as is customary with Redhat Enterprise Linux. This is documented here: http://rhn.redhat.com/errata/RHSA-2010-0542.html . At the same time, there are a number of other Redhat updates we would like to install on the two BDII servers, including this kernel update: http://rhn.redhat.com/errata/RHSA-2010-0504.html . We will be
making use of our DNS round-robin setup between the two servers to ensure that one of them will always be up and responding to queries.

Timekeeping Improvements on Virtual Machines

The unreliability of the timekeeping on all GOC virtual machines has been a cause for concern for some time. We have followed VMware's best practices (http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1006427) in terms of ntpd parameters, but this has not been enough, despite VMware's assurances that the "divider=10" kernel parameter is no longer necessary as of RHEL 5.3. After extensive testing we have determined that certain kernel command-line parameters will greatly improve the timekeeping, but in order to make use of them a virtual machine must be rebooted. We will therefore be systematically rebooting GOC services that reside on VMs after altering the kernel parameters to improve the timekeeping. We will also be taking advantage of the maintenance window to apply all recent Redhat security updates. This will affect all GOC services other than BDII (which does not reside on virtual machines), but downtime will in all cases be limited to under 15 minutes, and very likely under 5 minutes.