I have the following data that I'm trying to set up the props.conf file for:
--------
14/10/13 11:18:02 Start process
--------
Oracle Web Cache 11g (11.1.1.6)
Copyright (c) 1999, 2013, Oracle and/or its affiliates. All rights reserved.
[2014-10-13T11:18:02-06:00] [webcache] [ERROR:32] [WXE-13134] [config] [ecid: ] Oracle Web Cache did not find a matching CACHE element in webcache.xml for the current host name (l14172), Oracle instance (/opt/oracle/Middleware/Oracle_WT1/instances/instance1) and system component name (webcache1).
[2014-10-13T11:18:02-06:00] [webcache] [ERROR:1] [WXE-09000] [main] [ecid: ] Oracle Web Cache process of ID 23590 exits with code 1 at line 7743 of file loadcfg.c [label: Build 11.1.1.7.0 130113.0721]
[2014-10-13T11:18:02-06:00] [webcache] [NOTIFICATION:1] [WXE-08514] [logging] [ecid: ] Cache server process ID 23590 is shutting down.
--------
14/10/13 11:18:04 Start process
--------
Oracle Web Cache 11g (11.1.1.6)
Copyright (c) 1999, 2013, Oracle and/or its affiliates. All rights reserved.
[2014-10-13T11:18:04-06:00] [webcache] [ERROR:32] [WXE-13134] [config] [ecid: ] Oracle Web Cache did not find a matching CACHE element in webcache.xml for the current host name (l14172), Oracle instance (/opt/oracle/Middleware/Oracle_WT1/instances/instance1) and system component name (webcache1).
[2014-10-13T11:18:04-06:00] [webcache] [ERROR:1] [WXE-09000] [main] [ecid: ] Oracle Web Cache process of ID 23592 exits with code 1 at line 7743 of file loadcfg.c [label: Build 11.1.1.7.0 130113.0721]
[2014-10-13T11:18:04-06:00] [webcache] [NOTIFICATION:1] [WXE-08514] [logging] [ecid: ] Cache server process ID 23592 is shutting down.
The line breaking that I want to employ is to break at the "--------" line that has the year following it, OR the lines that begin with a square bracket ( "[" ). I have tried my REGEX in a couple of regex tools and they work in those, but not in Splunk. The config I'm using is:
[ unifier_webcache ]
SHOULD_LINEMERGE=true
BREAK_ONLY_BEFORE=^(--------[\n\r]1)|^(\[)
NO_BINARY_CHECK=true
It finds the line breaks that start with the square bracket, but not the ones that start with the hyphens.
Any clues as to how I can get it to do the proper line breaks?
Try with this
[unifier_webcache]
BREAK_ONLY_BEFORE = ^(\s*\d{2}/\d{2}|\[\d{4}-\d{2})
MAX_TIMESTAMP_LOOKAHEAD = 150
NO_BINARY_CHECK = 1
SEDCMD-removehyphens = s/--------//
SHOULD_LINEMERGE = true
Use this:
BREAK_ONLY_BEFORE=(^\[)|(^--------[\n\r]+\d\d)
The key is to add the plus quantifier, I'm guessing your line breaks are \r\n
which doesn't match plain old [\r\n]
, that's just one character.
Try with this
[unifier_webcache]
BREAK_ONLY_BEFORE = ^(\s*\d{2}/\d{2}|\[\d{4}-\d{2})
MAX_TIMESTAMP_LOOKAHEAD = 150
NO_BINARY_CHECK = 1
SEDCMD-removehyphens = s/--------//
SHOULD_LINEMERGE = true
The SEDCMD was the real trick to fixing the problem. I don't like having the data look different in Splunk from the log, but it doesn't work right otherwise, so it has become my solution.
BTW, the BREAK_ONLY_BEFORE didn't match the timestamp, but I had that anyway.
Thanks so much!