Exadata: 7 Useful Commands to check Port/Sensor Alarms

Hello all!

This days I had an alarm with message below:

Message=The aggregate sensor /SYS/CABLE_CONN_STAT has a fault.

There is some useful commands I used to verify all ports/sensors in my exadata cluster.

In summary, these commands:
1) Use Intelligent Platform Management Interface (IPMI) to read the Sensor Data Record (SDR) repository
2) Use Intelligent Platform Management Interface (IPMI) to view the ILOM SP System Event Log (SEL)
3) Display all host nodes with ibhosts
4) Use ibcheckstate to scan InfiniBand fabric and validate the port logical and physical state
5) Use ibcheckerrors to scan InfiniBand fabric and validate the connectivity as described in the topology file
6) Checking for sensor healthy from switch
7) Check the overall health of the InfiniBand switch, on the Exadata switch itself

The Commands are:

Continue reading

Advertisements

Oracle ASR: Communication Issues

Hello all,
Are you having notifications like this one from you ASR?

ALERT: Oracle Auto Service Request (ASR) has detected a heartbeat failure for these assets.

[list with "Serial"; "Hostnam"; "Information" of affected targets]

IMPACT: ASR would not be able to create a Service Request (SR) if a fault were to occur.
ACTION: Determine why the heartbeat has failed for these assets and resolve the issue.

This is only a notification saying that ASR not able to reach a target.
For detailed information on how to troubleshoot, you can access MOS Oracle Auto Service Request (ASR) No Heartbeat Issue – How to Resolve (Doc ID 1346328.2)

In general why, things to test are:
– Access from ASR server to transport.oracle.com, via https, using port 443.
– Access from ASR server to ASR assets, via http, using port 6481.

Continue reading

OEM Alarm – %MB of Audit Trail files (sizeOfOSAuditFiles:FILE_SIZE)

Hello All,
After upgrading a OEM to 13c, I started to receive notifications for event “sizeOfOSAuditFiles:FILE_SIZE“.

This is a new event implemented on OEM DB Plugin 12.1.0.7.0 under “Operating System Audit Records” metric group. Upgrading DB Plugin was part of OEM Upgrade change once we had some old versioned.

This event is only a notification related to file size for space management ends. The default thresholds are 10MB (warning) and 20MB (critical), which in most of times it’s a pretty low value.
This is specifically related to location under parameter audit_file_dest if you want to check.

Between options to reduce the noise are disable this metric or increase thresholds accordingly, which was what I did.
At this moment, I just increased thresholds to 500MB/2048MB, which I consider good values for the environment.

Some reference about can be found at:
– Enterprise Manager Oracle Database Plug-in Metric Reference Manual (Plug-in Release 12.1.0.7) – Database Instance – Operating System Audit Records
– EM 12c, EM 13c: Troubleshooting Database Metrics in Enterprise Manager 12c and 13c Cloud Control (Doc ID 2032156.1)

Hope it helps,
Cheers!

OEM Alarms: TaskZombieException

Hello all,
Recently I started to have several alarms like this:

Internal error detected: java.lang.Throwable:oracle.sysman.gcagent.tmmain.execution.LongOpManager$ZombieDetection:1017.

Or

Internal error detected: oracle.sysman.gcagent.task.TaskZombieException:oracle.sysman.gcagent.task.TaskFutureImpl$WrappedTask:620.

In agent log the message:

# $AGENT_INST/sysman/log/gcagent.log:

2017-06-13 12:54:09,232 [355:GC.Executor.14 (oracle_database:DB_DB12:%DB%)] ERROR - oracle_database:DB_DB12:%DB% oracle.sysman.gcagent.task.TaskZombieException: task declared as a zombie
at oracle.sysman.gcagent.task.TaskFutureImpl$WrappedTask.accountedCall(TaskFutureImpl.java:620)
at oracle.sysman.gcagent.task.TaskFutureImpl$WrappedTask.call(TaskFutureImpl.java:643)
at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303)
at java.util.concurrent.FutureTask.run(FutureTask.java:138)

To avoid generating pages and new incidents, I changed parameter _zombieCreateIncident to false on agents in related servers, as per described in MOS EM12c: Incident constantly raised for Oracle.sysman.gcagent.task.TaskZombieException: task declared as a zombie (Doc ID 2116834.1) regarding MOS Bug 22674258 – Zombie processes created for DB metric collection /workaround do not work.
I also added some other parameters that should help, like increasing the wait, all mentioned in same MOS notes.

Like this, in $AGENT_INST/sysman/config/emd.properties:

#GrepOra magical fixes:
_zombieSuspensions=true
_canceledThreadWait=900
_zombieThreadPercentThreshold=0
_zombieCreateIncident=false

And it solved the issue. At leat, the noise stopped. 🙂

Hope it helps,
Cheers!

OEM 12c+: Prevent OMS from Sending Old Notifications for Events

Are you receiving old notifications from OEM? Like 2 or 3 days past, mostly already solved or after a blackout?
Yeah, this is annoying, specially when getting floods and floods of notifications.

Ok, so here go a very good tip: You can set grace period for notifications! 🙂
Easy easy, do this way:

cd /bin
emctl set property -name oracle.sysman.core.notification.grace_period -value [provide value in minutes]

The oracle.sysman.core.notification.grace_period OMS parameter has been introduced in 12c and allows the user to configure the grace period within which the notification should be sent. The value is set in minutes.

For example:

emctl set property -name oracle.sysman.core.notification.grace_period -value 1440

With this, OMS sends only those notifications which have been raised in the past 1440 mins (last 24 hours) and ignores all the notifications for events / incidents created prior to this time period.

After this, you’ll need to start OMS:

emctl start oms

The oracle.sysman.core.notification.grace_period parameter applies to all the Notification methods, but if the requirement is to specify the grace period for a particular notification method only, you can use the below parameters accordingly:

oracle.sysman.core.notification.grace_period_connector: For Connectors
oracle.sysman.core.notification.grace_period_email: For email notifications
oracle.sysman.core.notification.grace_period_oscmd: For OS Command notifications
oracle.sysman.core.notification.grace_period_plsql: For PLSQL notifications
oracle.sysman.core.notification.grace_period_snmp: For SNMP Trap notifications
oracle.sysman.core.notification.grace_period_ticket: For ticketing tools

This is weel described as per MOS: 12c Cloud control: How to Prevent OMS from Sending out Old Notifications for Events / Incidents Occurred in the Past? (Doc ID 1605351.1)

Cheers!

EMagent 12c with high number of threads causing “su: cannot set user id: Resource temporarily unavailable”

Hello all,
Yeah, these days I got some errors and when validating the server found the fllowing error:

su: cannot set user id: Resource temporarily unavailable

As you can imagine, in order to fix the issue, I adjusted the /etc/security/limits.conf increasing oracle nprocs to:

oracle           soft    nproc           4047
oracle           hard    nproc           20384

Ok, turns out that after a while I got the same errors again…
After some investigating I find that the EM Agent process was with 5020 threads!
Take a look:

Continue reading

Metric evaluation error start – Get dynamic property error,Dynamic Category property error,Computation of a critical dynamic property failed. Retries Completed.

Hi all,
Getting this error? OMS 12.1.0.3?
Ok… After an investigation on MOS and also a SR opened with Oracle is was pointed to relation with Bug 17575631 CLUSTER DB HOME PAGE ERRORS- REGIONS FETCH REAL TIME DATA WILL NOT BE DISPLAYED on Database Plugin.
It’s solved in OMS 12.1.0.4 and also in 12.1.0.5.1 EM Database Plugin Bundle patch.

 I’ll show you how to workaround it. 🙂

Continue reading