Okay, sorry for the delay... I think I've figured this out. Are you able to build Heka from source? If so, maybe you could try out the `logstreamer_rotate_fix` branch that I've just pushed and see if it resolves your issue?
Thanks, -r On 03/29/2015 10:59 AM, rajive kumar wrote:
Hi Thank you... Best Regards RajiveKumar On 28 March 2015 at 12:14, Rob Miller <[email protected] <mailto:[email protected]>> wrote: I'm able to reproduce the issue, I'll work on getting a fix on the versions/0.9 branch over the next couple of days. -r On 03/27/2015 09:33 PM, rajive kumar wrote: Hi Guys Any idea on this.. Best Regards, On 25 March 2015 at 11:29, rajive kumar <[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>>> wrote: Hi Please find the link for the toml file and sh script which i have used to test. http://pastebin.ca/2963610 - toml file http://pastebin.ca/2963598 - sh script for simulate the log rotation. Steps I have followed 1. Run the script for some time by increasing the no of records 2. Start the heka while script is running. 3. After some time get the line count of generated files and heka output files ( Heka out put file have more than generated file count i.e its containing duplicate values ) Thank you in advance for the help. Best Regards, On 25 March 2015 at 05:40, Rob Miller <[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>>> wrote: If you're using the translation map as I've specified and you're still seeing the same file (or files) being read twice, then there might be a bug. I can't quite tell from what you're saying whether or not you've applied the change that I recommended. Can you paste your entire Heka config in a pastebin somewhere so I can look at it? Or at the very least the entire LogstreamerInput section, including the translation map subsection if you've applied it. Thanks, -r On 03/24/2015 12:07 PM, rajive kumar wrote: ---------- Forwarded message ---------- From: *rajive kumar* <[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>> <mailto:[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>>>> Date: 25 March 2015 at 00:35 Subject: Re: [heka] Logstreamer Input issue with current log and rolling file To: [email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>> <mailto:[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>>> Hi Thank you.. What you have mention is correct.I have tested that too. But my problem is when current file rolling Heka is reading all the rolled file again. For static file its working perfectly. For example In server log files line count is : 3015 But heka file reading log side is : 23065 file format which I have used is [smpp_logs] type = "LogstreamerInput" log_directory = "/media/rajive/Office/tmp/____test/sender/smpp" file_match= 'HS_06_SMPP_FULL_CDR(_(?P<Seq>____\d*))?' priority =["Seq"] [smpp_logs.translation.Seq] missing = 9999999999999 Please find the attachment for more details. Also I have started freshly i.e before start i have removed all the files. On 25 March 2015 at 00:08, Rob Miller <[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>> <mailto:[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>>>> wrote: You need to use what we call a "translation map" to map a missing `Seq` value to the highest possible value. The following is an example of what should work: [smpp_logs] type = "LogstreamerInput" log_directory = "/path/to/smpp/logs" file_match= 'smpp_cdr(_(?P<Seq>\d*))?' priority =["Seq"] [smpp_logs.translation.Seq] missing = 9999999999999999 Also be aware that you can test out the order that LogstreamerInput will load log files before you actually run Heka by pointing the heka-logstreamer utility at your configuration: hekad.readthedocs.org/en/v0.9.______0/pluginconfig/__logstreamer.____html#__verifying-settings <http://hekad.readthedocs.org/en/v0.9.____0/pluginconfig/logstreamer.____html#verifying-settings> <http://hekad.readthedocs.org/__en/v0.9.__0/pluginconfig/__logstreamer.__html#verifying-__settings <http://hekad.readthedocs.org/en/v0.9.__0/pluginconfig/logstreamer.__html#verifying-settings>> <http://hekad.readthedocs.org/____en/v0.9.0/pluginconfig/____logstreamer.html#verifying-____settings <http://hekad.readthedocs.org/__en/v0.9.0/pluginconfig/__logstreamer.html#verifying-__settings> <http://hekad.readthedocs.org/__en/v0.9.0/pluginconfig/__logstreamer.html#verifying-__settings <http://hekad.readthedocs.org/en/v0.9.0/pluginconfig/logstreamer.html#verifying-settings>>> -r On 03/24/2015 11:26 AM, rajive kumar wrote: Hi Guys Any idea on this I using LogstreamerInput to read my ongoing logs. Currently my logs has been rolling for every 15h min. When using heka its reading duplicate values. I.e its reading same file again and again. Following is the file pattern smpp_cdr (This the current log file) smpp_cdr_2015032416150055 smpp_cdr_2015032416300056 smpp_cdr_2015032416450057 I using following pattern to identify the file pattern file_match= 'smpp_cdr(_(?P<Seq>\d*))?' priority =["Seq"] This is read the rolled files perfectly but its reading the current log file (smpp_cdr) first. Also when current file gets rolled heka reading all the file again. Please advice / help on this. Thank you On 20 March 2015 at 18:44, rajive kumar <[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>> <mailto:[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>>> <mailto:[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>> <mailto:[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>>>>__> wrote: Hi Currently I am using Logstreamer input to process another applications current log file. Its working fine but I have an issue when file rolling -> its reading the file twice. Please help on this. Below the file structure I have in that system *trans_log* <- current log *150320001_trans_log * *150320002_trans_log * *150320003_trans_log * *150320004_trans_log * *150320005_trans_log * *150320006_trans_log* Note my files are rolling in every 15 mins and I want to read the previous logs and current log while its printing(i.e logging). Best Regards, _____________________________________________________ Heka mailing list [email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>> <mailto:[email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>>> https://mail.mozilla.org/______listinfo/heka <https://mail.mozilla.org/____listinfo/heka> <https://mail.mozilla.org/____listinfo/heka <https://mail.mozilla.org/__listinfo/heka>> <https://mail.mozilla.org/____listinfo/heka <https://mail.mozilla.org/__listinfo/heka> <https://mail.mozilla.org/__listinfo/heka <https://mail.mozilla.org/listinfo/heka>>> ___________________________________________________ Heka mailing list [email protected] <mailto:[email protected]> <mailto:[email protected] <mailto:[email protected]>> https://mail.mozilla.org/____listinfo/heka <https://mail.mozilla.org/__listinfo/heka> <https://mail.mozilla.org/__listinfo/heka <https://mail.mozilla.org/listinfo/heka>> _________________________________________________ Heka mailing list [email protected] <mailto:[email protected]> https://mail.mozilla.org/__listinfo/heka <https://mail.mozilla.org/listinfo/heka>
_______________________________________________ Heka mailing list [email protected] https://mail.mozilla.org/listinfo/heka
