Bogus jump second disrupts Linux systems
During a night of 31 Jul to 1 August, several servers that yield time information around NTP (Network Time Protocol) wrongly announced that clients should request a jump second. On Tuesday evening, Marco Marongiu pointed to this issue on one of a NTP project’s mailing lists. Now, reports from users whose systems practical a jump second during 00:00 Coordinated Universal Time (UTC) – 2am CET – can be found in places such as a Mythtv forum, on Twitter, on Google+ and on a NTP project’s mailing lists.
Unlike during a prior spin of a month, no leap second was scheduled to be practical final night. Occasionally added during a finish of Jun or December, such additional seconds are introduced to safeguard that Coordinated Universal Time never deviates from UT1 astronomical time by some-more than 0.9 seconds; this is designed to keep time time tighten to a time of day (i.e. a tangible position of a sun).
It is as nonetheless misleading because several NTP servers announced that a jump second would be combined final night; in a posting on one of a NTP project’s mailing lists, Marco Marongiu speculated that it could have been a “rather imaginative” Denial of Service (DoS) conflict on a tellurian scale.
If this is a case, a conflict substantially targeted badly confirmed Linux computers. A month ago, a genuine jump second triggered a bug in a Linux heart that crashed countless Linux systems; some were even disrupted to such an border that a processor finished adult operative underneath full load, wasting energy until a complement was rebooted or an director intervened. The heart developers analysed a problem and bound it during a growth of Linux 3.5; shortly afterwards, a corrections were incorporated into Linux heart versions 3.0.38, 3.2.24 and 3.4.6, that were all expelled in a second half of July. Although several Linux distributors also deployed a heart refurbish to repair a means of a problem, many systems are substantially still using exposed Linux kernels.
(fab)