running out of rootdbs space
joe reported sysadmin tables ph_run and ph_alert growing to about 5 million rows (2GB of rootdbs in a month) after moving from IDS 11.5FC5 back to FC3, with Auto Update Statistics Refresh errors -23197 and -640 in the log. Roy Mercer advised moving the sysadmin database to another dbspace. theBP asked how the downgrade was done. Unresolved.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
After migrating to ids11.5FC5 and back to old versio 11.5FC3 due to
some bug.
database 'sysadmin' table : ph_run & ph_alert
grow very fast, 2GB was empty in one month.
database sysadmin;
select tabname,nrows from systables ;ph_run & ph_alert has 5 millions rows .
select * from ph_alert :id 6300
alert_task_id 19
alert_task_seq 24
alert_type WARNING
alert_color GREEN
alert_time 2010-01-02 01:00:26
alert_state NEW
alert_state_chang+ 2010-01-02 01:00:26
alert_object_type MISC
alert_object_name
alert_message TASK NAME Auto Update Statistics Refresh
LOCATION <U
NKNOWN>
ERROR -23197 Database locale information mi
smatch.
alert_action_dbs sysadmin
alert_action
A lots of Garbage insert continuously .
What happen with these ?
Can I delete these two tables ?
Look in the manual, there is a command to move the sysadmin database
into a different DBSPACE.
Once you get the database moved out of rootdbs, you can then figure
out what is filling up the tables.
Roy
On Jan 21, 1:04 pm, joe <joecl...@gmail.com> wrote:
> After migrating to ids11.5FC5 and back to old versio 11.5FC3 due to
> some bug.
>
> database 'sysadmin' table : ph_run & ph_alert
> grow very fast, 2GB was empty in one month.
>
> database sysadmin;
> select tabname,nrows from systables ;> ph_run & ph_alert has 5 millions rows .
>
> select * from ph_alert :> id 6300
> alert_task_id 19
> alert_task_seq 24
> alert_type WARNING
> alert_color GREEN
> alert_time 2010-01-02 01:00:26
> alert_state NEW
> alert_state_chang+ 2010-01-02 01:00:26
> alert_object_type MISC
> alert_object_name
> alert_message TASK NAME Auto Update Statistics Refresh
> LOCATION <U
> NKNOWN>
> ERROR -23197 Database locale information mi
> smatch.
> alert_action_dbs sysadmin
> alert_action
>
> A lots of Garbage insert continuously .
> What happen with these ?
> Can I delete these two tables ?
One more things,online.log's messages list :
02:29:35 SCHAPI: Error -640 QPlan sanity failure (3778.105).
02:29:35 SCHAPI: Type: TASK, Name: Auto Update Statistics Refresh,
Location: NULL.
02:29:35 SCHAPI: Type: TASK, Name: Auto Update Statistics Refresh,Error Message: Unable to execute function [ aus_refresh_stats
(integer,integer) ] .
What made 'aus_refresh_stats' unable to execute, and keep
writing error log ?
On 21/01/2010 18:04, joe wrote:
> After migrating to ids11.5FC5 and back to old versio 11.5FC3 due to
> some bug.
>
> database 'sysadmin' table : ph_run& ph_alert
> grow very fast, 2GB was empty in one month.
>
> database sysadmin;
> select tabname,nrows from systables ;> ph_run& ph_alert has 5 millions rows .
>
> select * from ph_alert :> id 6300
> alert_task_id 19
> alert_task_seq 24
> alert_type WARNING
> alert_color GREEN
> alert_time 2010-01-02 01:00:26
> alert_state NEW
> alert_state_chang+ 2010-01-02 01:00:26
> alert_object_type MISC
> alert_object_name
> alert_message TASK NAME Auto Update Statistics Refresh
> LOCATION<U
> NKNOWN>
> ERROR -23197 Database locale information mi
> smatch.
> alert_action_dbs sysadmin
> alert_action
>
> A lots of Garbage insert continuously .
> What happen with these ?
> Can I delete these two tables ?
>
>
>
>
>
Just out of interest ...
Migration to 11.50.FC5 from ????, presumably just starting the original engine under 11.50.FC5.
Reverting back to 11.50.FC3 ... Did you do a restore? Did you just shutdown under 11.50.FC5 and startup under 11.50.FC3? Did you run
onmode -b?