300 GB database export should not be taking 2 days. What kind of servers you have? You can try making your export faster.
You can try exporting at schema level. start multiple export sessions at the same time. This is a sort of parallel processing your export. Make sure your export dump files are not being written on same physical disk or any NFS mounted disk.
How can one improve Import/ Export performance?
For public synonym, you can make your own scripts using SELECT DBMS_METADATA.GET_DDL or use any third party tool like TOAD, SQL developer etc and just run that in a new database. For synonyms:
select 'create public synonym ' || table_name || ' for ' || table_owner || '.' || table_name || ';' from dba_synonyms
where owner='PUBLIC' and table_owner not in ('SYS', 'SYSTEM')
order by table_owner;
You can also TRACE the export session to see where it is stuck.
Thank you all
Sorry for using the wrong word its not stuck its too slow. The exp command is FULL=y and 10 dumpfiles with 30GB each. mount point is NFS .. OS SunOS 5.8 it is still running cant post entire details i mean log contents and all since its prod please bear with me.
I know NFS is really slow but i dont have any other mount points with enough space
It ran for 3 days total .Is this all because of "NFS" drive because thats only thing we can blame as of now but if we get space in
local drives and if it slows again then all fingers would be pointed to us. I have asked for 300GB space in local file system will increase buffer size and give direct=y
and statistics=none i am guessing that would work any thoughts on that. Any suggestion or experience with slowness
I had started a different discussion since the question was different i got this reply from dean i had asked if starting multiple exports of schemas serve same purpose as full=y
"If you need the complete contents of the database, then you need to do full=y. Some objects are not owned by a schema so they would be be exported in a schema mode export.
I'm not sure if there is a way to do what you want.
I think just by moving from NFS to local should do the trick and setting all other parameters is a plus. I think even if you just export your all the schemas, you will be good. I had already given you a query to copy all PUBLIC objects.
Also there used to be a BUG which used to skip the jobs created with DBMS_JOB.
But after you are done with the IMPORT, you can always compare your old and new databases by using any third party tool such as SQL developer or TOAD etc.