Requirement:
There were couple of TAC cases that were open for system uptime being reset after 496 days of uninterrupted service.
The system uptime counter on 2910 switches running a W.15.14.0009 build roll over after approximately 496 days of uptime.
Solution:As the definition explains, the counter is measured in hundredths of a second, so after how many days will the counter start on 0 again?
4,294,967,295 100s/sec = 42,949,672.95 seconds
42,949,672.95 seconds = 715,827.8825 minutes
715,827.8825 minutes = 11,930.464708334 hours
11,930.464708334 hours = 497.10269618055555555555555555556 days
So, after 497.10 days of uninterrupted uptime, the system uptime counter will roll-over and start again on 0.
Configuration:The sys Uptime object that’s behind the uptime counter comes from an IETF RFC and has the following definition:
>snmptranslate -m all -TB -Td -On sysUptime
.1.3.6.1.2.1.1.3
sysUpTime OBJECT-TYPE
-- FROM SNMPv2-MIB, RFC1213-MIB
SYNTAX TimeTicks
MAX-ACCESS read-only
STATUS current
DESCRIPTION "The time (in hundredths of a second) since the network management portion of the system was last re-initialized."
This is a 32-bit counter. As such, the maximum value is 232 -1 = 4,294,967,295. Once the counter has reached that value, it will roll-over and start on 0 again, as specified in the IETF RFCs.
VerificationIn relation to the issue statement, the following observations are then expected behaviour:
1) The system uptime is reset after approximately 496 days of uninterrupted uptime. Per explanation above, this is correct. This is also valid for any vendor and any system that has a 32-bit system uptime counter.
2) The counters are not reset after a switch has been rebooted. Correct. The counters work fine and will only roll-over to 0 after 497.10 days of uninterrupted uptime. By rebooting the switch, the counter is reset and will start at 0 again.