On 10/24/2011 06:38 PM, Merlin Moncure wrote:
On Mon, Oct 24, 2011 at 3:22 AM, Manoj K P<ma...@comodo.com> wrote:Postgres recovered from data folder , after that some queries shows errorselect * from table2 order by app_id ; - its work ( 50000000 data) select * from table2 order by app_id desc ; - its work Here app_id contains binary index select * from table2 order by id ; - its work ( 50000000 data) select * from table2 order by id desc ; - shows following error manoj[local] postgres ERROR: index "tbl2_id_pkey" contains unexpected zero page at block 311121 manoj [local] postgres HINT: Please REINDEX it. index "tbl2_id_pkey" is primary key Reindex not pratical this table (250 GB data) How i can solve this ?It looks like you have data corruption. Your first priority should be to do the minimum necessary to get a usable database dump and then to get back online. I would consider your database off line until this is done. Essentially, the only practical ways to repair the error you see would be to reindex or restore from backup. Did you have any unexpected power events? Server crashes? merlin
Actually this is backup server ,no power failure& Server crashes happens in between , The data folder is base backup(|pg_start_backup)| on the server. pg_dump not practical in this case because of data size. After recovering from base backup run WAL recovery file successfully and its updates current date data. When i am running query on big table (more than 5 Million data) shows error other wise its work fine. select id , name from table2 order by id ; It uses the index select id , name from table2 order by id desc ; - shows the error Error happens only using (DESC clause against primary key) , all other case its work fine REINDEX will take ages -- *Manoj K P* *Postgres DBA* *Comodo India*
smime.p7s
Description: S/MIME Cryptographic Signature
