Wikipedia

Search results

Friday 20 June 2014

Archive destination gets full in Prod

Issue Description :
-------------------

For the past 3 days, The early morning production instance gets down and alert.log shows that archive destination gets full and APPS_TS_DATA and APPS_TS_IDX tablespace also gets full.
Within a day, the 500GB archive destination was full.

Solution :
----------

I have raised an SR for this issue and have collected the information about the session which was consuming more.

1. I have generated AWR report using following method :

conn system/manager
@$ORACLE_HOME/rdbms/admin/awrrpt.sql
It prompts for
1. html
2. start snapshot id - This should be before issue raise (There is no database down happened between the IDs)
3. end snapshot id - This should be after issue raised
4. report filename - User defined

This report completely shows (especially,Segments by Physical Writes)

the majority of the block changes (= creating the redo) during this time where done on

XLA.XLA_DIAG_SOURCES TABLE in APPS_TS_TX_DATA tablespace
and it's index
XLA.XLA_DIAG_SOURCES_N1 INDEX in APPS_TS_TX_IDX tablespace

2. SR support team has given the correct solution

As per Note 465834.1, the XLA_DIAG_SOURCES table keeps diagnostic data, once the diagnostic data has been given to support on an issue, there is no reason to keep this information. Also, diagnostic information can be regenerated as well if needed.

It is fine to truncate this table:
SQL> truncate table xla.XLA_DIAG_SOURCES

Perhaps you should also consider disabling SLA diagnostics.

We have truncated the table on PROD and set "SLA: Enable Diagnostics" to No.

Now issue got resolved...
Happiest day will be tomorrow...

No comments:

Post a Comment