Re: SLRUs in the main buffer pool - Page Header definitions

Hi, 

We have been working on adding page headers to the SLRU pages, as part of the migration for SLRU to buffer cache. We’ve
incorporatedThomas Munro’s patch and Heikki’s Storage manager changes[1] and have a patch for early feedback. 
 

As part of our changes we have:

1. Added page headers to the following
  
               *Commit_TS
      *CLOG
      *MultiXact
      *Subtrans
      *Serial (predicate.c)
      *Notify (async.c)

For commit_ts, clog and MultiXact, the PageXLogRecPtr field is populated with the LSN returned during the creation of a
newpage; as there is no WAL record for the rest, PageXLogRecPtr is set to “InvalidXlogRecPtr”.
 

There is one failing assert in predicate.c for SerialPagePrecedes with the page header changes; we are looking into
thisissue.
 

The page_version is set to PG_METAPAGE_LAYOUT_VERSION (which is 1)


2. Change block number passed into ReadSlruBuffer from relative to absolute, and account for SLRU’s 256kb segment size
inmd.c.
 



The changes pass the regression tests. We are still working on handling the upgrade scenario and should have a patch
outfor that soon.
 

Attached is the patch with all changes (Heikki and Munro’s patch and page headers) consolidated 


Thanks,
Rishu Bagga, Amazon Web Services (AWS)

[1] https://www.postgresql.org/message-id/128709bc-992c-b57a-7174-098433b7faa4@iki.fi

[2] https://www.postgresql.org/message-id/CA+hUKG+02ZF-vjtUG4pH8bx+2Dn=eMh8GsT6jasiXZPgVxUXLw@mail.gmail.com







On 9/27/22, 6:54 PM, "Bagga, Rishu" <bagrishu@amazon.com> wrote:

    Hi all,

    PostgreSQL currently maintains several data structures in the SLRU
    cache. The SLRU cache has scaling and sizing challenges because of it’s
    simple implementation. The goal is to move these caches to the common
    buffer cache to benefit from the stronger capabilities of the common
    buffercache code. At AWS, we are building on the patch shared by Thomas
    Munro [1], which treats the SLRU pages as part of a pseudo-databatabe
    of ID 9. We will refer to the pages belonging to SLRU components as
    BufferedObject pages going forward.

    The current SLRU pages do not have any header, so there is a need to
    create a new page header format for these. Our investigations revealed
    that we need to:

    1. track LSN to ensure durability and consistency of all pages (for redo
       and full page write purposes)
    2. have a checksum (for page correctness verification).
    3. A flag to identify if the page is a relational or BufferedObject
    4. Track version information.

    We are suggesting a minimal BufferedObject page header
    to be the following, overlapping with the key fields near the beginning
    of the regular PageHeaderData:

    typedef struct BufferedObjectPageHeaderData
    {
        PageXLogRecPtr pd_lsn;
        uint16_t       pd_checksum;
        uint16_t       pd_flags;
        uint16_t       pd_pagesize_version;
    } BufferedObjectPageHeaderData;

    For reference, the regular page header looks like the following:
    typedef struct PageHeaderData
    {
        PageXLogRecPtr    pd_lsn;
        uint16_t    pd_checksum;
        uint16_t    pd_flags;
        LocationIndex   pd_lower;
        LocationIndex   pd_upper;
        LocationIndex   pd_special;
        uint16_t           pd_pagesize_version;
        TransactionId   pd_prune_xid;
        ItemIdDataCommon  pd_linp[];
    } PageHeaderData;

    After careful review, we have trimmed out the heap and index specific
    fields from the suggested header that do not add any value to SLRU
    components. We plan to use pd_lsn, pd_checksum, and pd_pagesize_version
    in the same way that they are in relational pages. These fields are
    needed to ensure consistency, durability and page correctness.

    We will use the 4th bit of pd_flags to identify a BufferedObject page.
    If the bit is set then this denotes a BufferedObject page. Today, bits
    1 - 3 are used for determining if there are any free line pointers, if
    the page is full, and if all tuples on the page are visible to
    everyone, respectively. We will use this information accordingly in the
    storage manager to determine which callback functions to use for file
    I/O operations. This approach allows the buffercache to have an
    universal method to quickly determine what type of page it is dealing
    with at any time.

    Using the new BufferedObject page header will be space efficient but
    introduces a significant change in the codebase to now track two types
    of page header data. During upgrade, all SLRU files that exist on the
    system must be converted to the new format with page header. This will
    require rewriting all the SLRU pages with the page header as part of
    pg_upgrade.

    We believe that this is the correct approach for the long run. We would
    love feedback if there are additional items of data that should be
    tracked as well. Alternatively, we could re-use the existing page
    header and the unused fields could be used as a padding. This feels
    like an unclean approach but would avoid having two page header types
    in the database.



    [1] -
https://www.postgresql.org/message-id/flat/CA+hUKGKAYze99B-jk9NoMp-2BDqAgiRC4oJv+bFxghNgdieq8Q@mail.gmail.com



    Discussed with: Joe Conway, Nathan Bossart, Shawn Debnath


    Rishu Bagga

    Amazon Web Services (AWS)