I'm having an issue where the timestamp for an event is presented one hour in the past.
ex: for the following event produced on the splunk server:
20100831173739.374286
AvailableMBytes=1146
CommittedBytes=979374080
PagesPersec=12
PercentCommittedBytesInUse=21
wmi_type=Memory
i'm seeing the following time for the event on the web interface:
8/31/2010 4:37:39.3742
it should be 8/31/2010 5:37:39:372
some notes:
any thoughts?
Is the time zone set to GMT, or is the time zone set to "GMT: London,Dublin" or whatever the Windows interface calls it, or is it set to "GMT: Coordinated Universal Time"? During the summer, London is on British Summer Time, and the Windows UI is not clear that there is a difference between real GMT and London time.
Nevertheless there is probably a bug in Splunk in that Splunk indexing seems to think that your machine is on "real" GMT, when in fact it's on British Time which switches between GMT and BST, and so does not make the appropriate adjustment when it stores the timestamps internally. It's likely/possible that Splunk indexing suffers from that same confusion as Windows probably presents a confusing time zone name to the APIs as well. At the same time, when times are displayed to the user, they are probably correctly converted according to the current offset of the display timezone. Or maybe it's the other way around.
Is the time zone set to GMT, or is the time zone set to "GMT: London,Dublin" or whatever the Windows interface calls it, or is it set to "GMT: Coordinated Universal Time"? During the summer, London is on British Summer Time, and the Windows UI is not clear that there is a difference between real GMT and London time.
Nevertheless there is probably a bug in Splunk in that Splunk indexing seems to think that your machine is on "real" GMT, when in fact it's on British Time which switches between GMT and BST, and so does not make the appropriate adjustment when it stores the timestamps internally. It's likely/possible that Splunk indexing suffers from that same confusion as Windows probably presents a confusing time zone name to the APIs as well. At the same time, when times are displayed to the user, they are probably correctly converted according to the current offset of the display timezone. Or maybe it's the other way around.
Ah, then the problem is that Splunk probably just looks at the time zone name/description, and ignores the DST flag.
i had it set to GMT: Dublin, Edinburgh, Lisbon, London with the daylight saving flag turned off. i changed it to GMT: Monrovia, Reykjavik and that resolved my issue. Thanks,
Hi Atul,
can you please let em know what props u have used to resolve this issue.
My issue is something like that, the logs are coming one hour delay(index time) from the venet generated time. And I need to change it to GMT+1 for BST as we are in BST currently?
Thanks,
Ankit