This SAS note, intended to be informational only, documents the areas that SAS Fraud Management customers might want to review before the next official leap-second event. Nearly all modern operating systems assume that 1 day=24x60x60=86400 seconds in all cases. In Coordinated Univsersal Time (UTC), however, about once every year or two there is an extra second, called a "leap second." This event is used to synchronize the atomic and astronomical clocks.

The leap second is the last second of the day, and always on December 31 or June 30. For example, the last minute of the year 1995 was 61 second long, thanks to an added leap second. Most computer clocks are not accurate enough to be able to reflect the leap-second distinction. The addition of the leap-second is currently scheduled for June 30, 2015 at 23:59:59 where a 23:59:60 will be added to the UTC standard.

Be aware that the information contained here is not specifically addressed by SAS Fraud Management, and no issues have been identified in SAS^® applications. It is recommended that you review the information below to address your environments and take any questions or concerns to the appropriate third-party vendor, as detailed below.

Red Hat Enterprise Linux

For details, see Resolve Leap Second Issues in Red Hat Enterprise Linux.

Red Hat Enterprise Linux 4:

• You must have a kernel version. (It protects from a system crash.)
• Upgrade to kernel-2.6.9-80.EL or later.
• Verify your kernel version by submitting the following command:
  
  uname -r
• For additional information, see System hangs on printing the leap second insertion method.

Systems that are not synchronized by the Network Time Protocol Daemon (NTPD):

• You need to have the Time Zone database (tzdata) package version tzdata-2015a-1 or later.
• Verify your version of tzdata by submitting the following command:
  
  yum list tzdata
• For additional information , see tzdata enhancement update.

Red Hat Enterprise Linux 5:

• You must have a kernel version. (It protects from a system crash.)
• Upgrade to kernel-2.6.18-164.el5 or later./li>
• Verify your kernel version by submitting the following command:
  
  uname -r
• For additional information, see System hangs on printing the leap second insertion message.

Systems that are not synchronized by the Network Time Protocol Daemon (NTPD)

• You need to have the Time Zone database (tzdata) package version tzdata-2015a-1 or later.
• Verify your version by submitting the following command:
  
  yum list tzdata
• For additional information, see tzdata enhancement update.

Red Hat Enterprise Linux 6:

• You must have a kernel version. (It protects from a system crash.)
• RHEL 6.1.z EUS kernel 2.6.32-131.30.2.el6 or later
• RHEL 6.2.z EUS kernel 2.6.32-220.25.1.el6 or later
• RHEL 6.3 kernel 2.6.32-279.5.2.el6 or later
• RHEL 6.4 GA kernel 2.6.32-358.el6 or later
• Any application that uses International Atomic Time (TAI) should have the following kernel versions:
• RHEL 6.2.z kernel 2.6.32-220.62.1.el6 or later
• RHEL 6.4.z kernel 2.6.32-358.59.1.el6 or later
• RHEL 6.5.z kernel 2.6.32-431.56.1.el6 or later
• RHEL 6.6.z...This version is currently being worked on under internal BZ 1199134.
• It is recommended that you keep the Network Time Protocol (NTP) daemon updated.

Java

This impact of the leap second for Java depends on the host environment. Although the Date class is intended to reflect coordinated universal time (UTC), it might not do so exactly, depending on the host environment of the Java Virtual Machine.

For additional information, see Class Date.

AIX Operating Environment

AIX does not have - and never has had - any direct awareness of the leap second. There have been several leap-second events that have occurred throughout the history of AIX, but no known AIX issues have been reported. However, the Network Time Protocol (NTP) daemon is available under AIX for those who need precise time synchronization with other systems.

AIX systems can fall into one of three scenarios:

• Sub-second synchronization is not an issue; no NTP is in use. The system runs as normal, unaware that its own time is one second offset from where it had been previously.
• System time synchronization is needed. In this case, use NTP daemon. Your AIX system can run the NTP time service as a client, synchronizing to another system or an external time server (for example, time.nist.gov). The NTP client slowly adjusts the system clock to synchronize with the time server, thus accommodating the leap second.
• The AIX system is itself acting as a primary NTP master server, without being, in turn, a client of another time server. This scenario is rare and is not recommended if you need sub-second time synchronization. However, if you need to configure your system in this way, you can do so following the instructions at ntp.org (specifically, as of this writing, Configuring NTP).

Most systems fall under scenarios #1 or #2, and you do not need to take any action.

For more information, see Leap Second Information for AIX.

WebSphere MQ

When the leap second was added on June 30, 2012, it as subsequently observed that a WebSphere MQ queue manager that was running in the Linux operating environment, generated many first-failure symptom report (FDC - file delete child) files related to resource issues or constraints. These FDC files commonly reported rc=xecP_E_NO_RESOURCE.

When the leap second is added, your queue manager might stop responding or you might experience high CPU use. The FDC files are generated on a daily basis. These files might have probes of XY348010 or XC272003 from the xcsCreateThread process. However, there might be other FDC files with different probes as well.

The leap second at the end of June in 2012 caused several problems with otherwise unexplainable high CPU usage on Linux systems. WebSphere MQ does not directly make the calls that experience the problem. But it does use the pthreads library (NPTL), which in turn uses fast user space (futexes), which can experience this problem. Busy systems that run WebSphere MQ and other products are susceptible to this problem.

For details about this problem, see Anyone else experiencing high rates of Linux server crashes during a leap second day?

For additional information, see Resolve Leap Second Issues in Red Hat Enterprise Linux.

For steps to resolve this issue, see Leap Second may cause Linux to freeze.

IBM DB2 Database

Currently, DB2 is configured to read its timestamp from the operating environment in which DB2 is installed. You might want to check with your operating-system administrator to find out how your environment will handle the addition of the leap second.

For additional information, see 75 ways to demystify DB2: #11: Tech Tip: What is the impact of leap second on DB2.

Operating System and Release Information

Product Family	Product	System	SAS Release
			Reported	Fixed*
SAS System	SAS Fraud Management	64-bit Enabled AIX
		Linux
		Linux for x64

* For software releases that are not yet generally available, the Fixed Release is the software release in which the problem is planned to be fixed.

---

This note is for informational purposes to assist SAS Fraud Management customers in addressing any concerns that are related to the addition of the leap second (June 30, 2015 at the time of this note).

Type:	Usage Note
Priority:

Date Modified:	2015-06-03 09:38:36
Date Created:	2015-05-27 15:01:23