The Percentage Of Requests That Resulted In Errors Is …..%

Part of our ongoing support and tuning of OEM 12c, I was woken up this morning by our help desk asking me to take a look at a critical metric error the OHS1 within OEM12c.  Typically, I would just tell them to ignore the error and move on.  For some reason this morning, think it is the lack of sleep; I jumped right in and took a look.

With any OHS1 related messages, the first place that should be referenced is the “ohs1.log”.  This is the Oracle HTTP Server (OHS) log which is rolled over periodically and must be maintained disk space.  Purging of the log, once rolled over can be done using OS schedulers such as crontab or windows scheduler.

The OHS1 log file can be found at:

<EM_INSTANCE_BASE>/<webtier_instance_name>/diagnostics/logs/OHS/<ohs_name>

Example:

/oracle/app/product/12.1.0.2.0/middleware/gc_inst/WebTierIH1/diagnostics/logs/OHS/ohs1

What prompted this alert to arise are the following messages being logged (one or both messages) to the ohs1.log

[OHS] [ERROR:32] [OHS-9999] [core.c] [host_id: <host>] [host_addr: <host_address>] [tid: 25] [user: oracle] [ecid: <ecid>] [rid: 0] [VirtualHost: <virtual host>]  Invalid method in request 0

[OHS] [INCIDENT_ERROR:10] [OHS-501] [mod_context.c] [host_id: <host>] [host_addr:  <host_address>] [pid: <pid>] [tid: 23] [user: oracle] [VirtualHost: main]  yodContextUnwrap(,FALSE) failed

According to the MOS note 1495483.1, the cause of these messages cannot be identified, but similar false alerts can occur in 12.1.0.1 release of OEM (with or without BP1).  In this note, the solution suggested is to upgrade to 12.1.0.2 or higher of OEM.  Additional solutions are to contact support if alerts due to [mod_context.c] messages keep appearing in the ohs1.log or to increase the metric threshold for “Error Rate(%)” on the OHS1 target.

Since the OEM12c environment, I’m working with is already at 12.1.0.2; it is easier to increase the threshold for this metric than anything else.  Is this the correct approach? 

Implement solutions:

Go to target for OHS1 in OEM12c

Oracle HTTP Server -> Monitoring -> All Metrics

Once on the OHS1 home page, ensure that the “view by” drop down says “Metrics”.

In the tree below the drop down menu, select:

OHS Server Metrics -> Error Rate (%) -> Modify Thresholds

Update the threshold with a setting higher presently set.

Please follow and like: