Re: pageinspect: Hash index support - Mailing list pgsql-hackers

From Ashutosh Sharma
Subject Re: pageinspect: Hash index support
Date
Msg-id CAE9k0PkepznkYUfdMh=jcQXrdx43RXepRut0HgNitOuK1ZEreA@mail.gmail.com
Whole thread Raw
In response to Re: pageinspect: Hash index support  (Jesper Pedersen <jesper.pedersen@redhat.com>)
List pgsql-hackers
Hi,

> git apply w/ v4 works here, so you will have to investigate what happens on
> your side.
>

Thanks, It works with v4 patch.

> As these functions are marked as superuser only it is expected that people
> provides the correct input, especially since the raw page structure is used
> as the input.

Well, page_stats / page_items does accept bitmap page as input but
these function are not defined to read bitmap page as bitmap page
doesnot have a standard page layout. Infact if we pass bitmap page as
an input to page_stats / page_items, the output is always same. I
think we need to have a separete function to read bitmap page. And i
am not sure why should a server crash if i pass meta page to
hash_page_stats or hash_page_items.


> The "if" statement will need updating once the CHI patch goes in, as it
> defines new constants for page types.

Not sure if CHI patch is adding a new type of hash page other than
holding an information about split in progress. Basically my point was
can we have hash page of types other than meta page, bucket page,
overflow and bitmap page. If pageinspect tool finds a page that
doesnot fall under any of these category shouldn't it error out.

With Regards,
Ashutosh Sharma
EnterpriseDB: http://www.enterprisedb.com



pgsql-hackers by date:

Previous
From: Robert Haas
Date:
Subject: Re: more parallel query documentation
Next
From: Robert Haas
Date:
Subject: Re: Fix Error Message for allocate_recordbuf() Failure