Use InfoSphere CDC event
notifications to receive alerts when the latency exceeds certain defined
thresholds.
Notifications are most useful when you analyze
replication activities with IBM
InfoSphere CDC Management Console and
you want to detect events that are happening in the source datastores,
target datastores, and subscriptions.
Procedure
To set up InfoSphere CDC event
notifications:
- Determine the SMTP server name by consulting the network
administrator.
- Log in to the InfoSphere CDC Management
Console as access-server-user.
- Stop mirroring and end replication on the subscription
for your shadow tables. InfoSphere CDC does
not allow any configuration changes while a subscription is active.
- Click .
- Right-click the subscription, select , and click OK.
- Set email notification at the datastore-level:
- Click .
- Right-click the datastore for your shadow tables and
select Notifications.
- Click the Source tab.
- From Notification Categories,
select the first category for which you want to enable notifications.
- Select the INTERNET MAIL box,
and enter the following information:
- In the SMTP Mail Host in the hostname box, enter the name the
outgoing mail server.
- In the Alert Account box, enter the one or more email addresses
or a distribution list that you want to notify. If you specify multiple
email addresses, use a semicolon (;) to separate them.
- In the Sender Mail Address box, enter the email address.
- In the Sender Mail Password box, enter the password of the sender.
- Click Copy Setting.
- Select the notification categories to apply the same
notification settings. Because in shadow tables the source and the
target are the same datastore, select Target categories from
that list.
- Select the Target > Apply > Informational category
to copy the notification settings and to enable latency threshold
and notification for the subscription.
- Return to the main Configuration > Datastores view
by clicking OK in the Copy Notification
Settings window and then also in the Notifications window.
- Set the latency threshold and notification:
- Click .
- Right-click the subscription for your shadow tables
and select Latency Thresholds.
- Select the Notify when latency crosses these
thresholds check box.
- Specify a warning threshold in minutes. A good value
is 75% of the value of the CURRENT REFRESH AGE special register.
- Specify a problem threshold in minutes. The value must
be greater than the warning threshold. A good value is 90% of the
CURRENT REFRESH AGE.
- Click OK. Do not click Set
Notification, because you specified all the required notification
settings in step 4.
- Set event filtering to disable event notification for InfoSphere CDC event
IDs that you are not interested in.
Enabling Target Informational
events to capture latency threshold notifications returns notifications
for event IDs that you are might not interested in.
- Click .
- Right-click the subscription and select Notifications.
- In the Notification Categories window,
click the Target tab.
- Click Filter Messages. The event
ID filtering applies to all the categories of the target datastore.
- In the list, enter the event IDs that you do not want
to be notified of (one event on each line), and select the Do
not send these messages option.
- Filter the same event IDs for the source datastore,
because it is the same as the target datastore.
- Start mirroring to start receiving event notifications
by email.
- Click .
- Right-click the subscription name and select Start
Mirroring.
- In the Start Mirroring window,
specify the mirroring method by selecting Continuous,
and click OK.
The following text shows examples of event notifications that
are sent by email:
Event ID: 1463
Publisher(S)/Subscriber(T): S
System ID: SUBS_SHADOW_TABLE_LI_SVT_INST1
Event Text: +++ Subscription SUBS_SHA is starting in Continuous Mirroring
mode.
Event ID: 3172
Publisher(S)/Subscriber(T): T
System ID: SUBS_SHA
Event Text: Latency for journal JOURNAL has exceeded the problem threshold
of 15 minutes.