wrong datestamp in "show log"

[Joel_Leonhardt]

FYI, HP has released information and a fix for this bug; see below.

Also, additional symptoms noted in our environment (hp5412zl, K.15.07.0008):

- Upon  snmp sysUpTime reaching approx 4294965000 (hundredths of a second since restart), or 497 days, it dropped to zero and started climbing again; this would normally indicate a switch restart, so triggered a monitoring system alert.

- CLI command "show uptime" indicated the switch to have been up for only a few minutes.

- But other CLI commands (eg: "show span" with long time since last change, and "show int" with high packet counts) indicated the switch to have been up for an extended period.

- And the log file had the symptoms mentioned in this discussion, with previous logs all timestamped correctly, but new logs (since sysUpTime was reset), timestamped starting about Dec 27, 2012, 9am.

- This happened to all 3 of our hp5400 switches that had been up for 497 days.

HP's information on this issue:

**Version K.15.13.0004 Fix List

Event Log (CR_0000127436) - After the switch uptime reaches 497 days, the timestamp entries in the event log become erratic with gaps of several hours or days. In some cases the timestamps revert to previous months and years, even though SNTP updates with those wrong timestamps report the correct date and time.

Reference:

http://h20566.www2.hp.com/portal/site/hpsc/template.BINARYPORTLET/public/kb/docDisplay/resource.process/?spf_p.tpst=kbDocDisplay_ws_BI&spf_p.rid_kbDocDisplay=docDisplayResURL&spf_p.rst_kbDocDisplay=wsrp-resourceState%3DdocId%253Demr_na-c04084207-5%257CdocLocale%253D

[Gordon Cook]

I understand the fix for this issue is to upgrade the firmware.   We do plan to do that it is just takes a lot of planning since we have to shut down most of our servers to do this.  Is there a way to reset the time for the logs or something in the meantime.  Trying to troubleshoot is a real pain right now.  We have a E5412ZL