Done.
I followed the steps below after reconnecting the filenode:
Select * from "05122016".repositorio
Count -> 0
Then, I execute the commands:
Analyze "05122016".repositorio;
Count -> 0
Reindex "05122016".repositorio;
Count -> 1509046
Yes!
But...
select * from "05122016".repository where id_documento = '{A647885D-9738-4617-ACE6-6351F6ADD341}'
Returns the error below:
Missing chunk number 0 for toast value 10259186 in pg_toast_9277966
I tried to do the commands below, but the error persists:
REINDEX table pg_toast.pg_toast_9277966;
VACUUM ANALYZE pg_toast_9277966
VACUUM ANALYZE "05122016".repositorio;
REINDEX TABLE pg_statistic;
I read in the forum to run the command:
Delete from pg_statistic;
Reindex table pg_statistic;
Vacuum analyze;
Is it okay to delete the pg_statistic table?
2017-06-05 9:24 GMT-03:00 tel medola <tel.medola@xxxxxxxxx>:
Sorry. I got sick these days and could not read my emails.Thanks for your help.I'll try to point to the direct node and see what happens.2017-06-01 10:29 GMT-03:00 Adrian Klaver <adrian.klaver@xxxxxxxxxxx>:On 06/01/2017 03:47 AM, tel medola wrote:
Did you get any help with this?
/I formatted correctly before sending the email. Maybe you should ask yourself if the mail server did not remove the formatting./
I was talking about help with your relfilenode issue, I learned to deal with the formatting awhile ago.
/
/
Well the relpages, reltuples are estimated values that can be updated with an ANALYZE./
/
/I can not make analyze on a table whose filenode is pointing to another reference. The table is empty, just because the filenode does not point to the correct ID./
Understood, I was just saying that if you could get the table pointing at the right relfilenode the other values would be synced up with an ANALYZE.
At this point it is time to just try something. You have file level backups of the various backups, so you can restore that if something goes wrong, correct?
For 01052016.repositorio with current pg_class entry of relfilenode of 13741352, change that back to the old entry of 5214489.
--
Adrian Klaver
adrian.klaver@xxxxxxxxxxx
