Re: Fix premature xmin advancement during fast forward decoding

2025-04-28 Thread Masahiko Sawada
On Mon, Apr 28, 2025 at 2:20 AM Amit Kapila wrote: > > On Sun, Apr 27, 2025 at 12:33 PM Zhijie Hou (Fujitsu) > wrote: > > > > I also adjusted the commit message to mention the commit f49a80c4 > > as suggested by Amit. > > > > Pushed. Thank you! Regards, -- Masahiko Sawada Amazon Web Services:

Re: Fix premature xmin advancement during fast forward decoding

2025-04-28 Thread Amit Kapila
On Sun, Apr 27, 2025 at 12:33 PM Zhijie Hou (Fujitsu) wrote: > > I also adjusted the commit message to mention the commit f49a80c4 > as suggested by Amit. > Pushed. -- With Regards, Amit Kapila.

RE: Fix premature xmin advancement during fast forward decoding

2025-04-27 Thread Zhijie Hou (Fujitsu)
he commit message to mention the commit f49a80c4 as suggested by Amit. Best Regards, Hou zj v2-0001-HEAD-PG17-Fix-premature-xmin-advancement-during-fast-forwar.patch Description: v2-0001-HEAD-PG17-Fix-premature-xmin-advancement-during-fast-forwar.patch From 16ae28e3b3b4c10bd82ad679194325eb56aa3965 Mon Se

Re: Fix premature xmin advancement during fast forward decoding

2025-04-26 Thread Masahiko Sawada
On Sat, Apr 26, 2025 at 5:01 AM Amit Kapila wrote: > > On Sat, Apr 26, 2025 at 12:01 AM Masahiko Sawada > wrote: > > > > On Fri, Apr 25, 2025 at 4:42 AM Amit Kapila wrote: > > > > > > Can you think of any better ideas? > > > > No idea. Hmm, there seems no reasonable way to fix this issue for ba

Re: Fix premature xmin advancement during fast forward decoding

2025-04-26 Thread Amit Kapila
On Sat, Apr 26, 2025 at 12:01 AM Masahiko Sawada wrote: > > On Fri, Apr 25, 2025 at 4:42 AM Amit Kapila wrote: > > > > On Fri, Apr 25, 2025 at 10:46 AM Masahiko Sawada > > wrote: > > > > > > What I'm concerned about is the back branches. With this approach all > > > back branches will have such

Re: Fix premature xmin advancement during fast forward decoding

2025-04-25 Thread Masahiko Sawada
On Fri, Apr 25, 2025 at 4:42 AM Amit Kapila wrote: > > On Fri, Apr 25, 2025 at 10:46 AM Masahiko Sawada > wrote: > > > > What I'm concerned about is the back branches. With this approach all > > back branches will have such degradations and it doesn't make sense to > > me to optimize SnapBuildCo

Re: Fix premature xmin advancement during fast forward decoding

2025-04-25 Thread Amit Kapila
On Fri, Apr 25, 2025 at 10:46 AM Masahiko Sawada wrote: > > What I'm concerned about is the back branches. With this approach all > back branches will have such degradations and it doesn't make sense to > me to optimize SnapBuildCommitTxn() codes in back branches. > One possibility could be that

Re: Fix premature xmin advancement during fast forward decoding

2025-04-24 Thread Masahiko Sawada
On Thu, Apr 24, 2025 at 9:56 PM Amit Kapila wrote: > > On Fri, Apr 25, 2025 at 8:14 AM Zhijie Hou (Fujitsu) > wrote: > > > > On Fri, Apr 25, 2025 at 5:44 AM Masahiko Sawada wrote: > > > On Tue, Apr 22, 2025 at 12:06 AM Zhijie Hou (Fujitsu) > > > wrote: > > > > > > > > Hi, > > > > > > > > When an

Re: Fix premature xmin advancement during fast forward decoding

2025-04-24 Thread Amit Kapila
On Fri, Apr 25, 2025 at 8:14 AM Zhijie Hou (Fujitsu) wrote: > > On Fri, Apr 25, 2025 at 5:44 AM Masahiko Sawada wrote: > > On Tue, Apr 22, 2025 at 12:06 AM Zhijie Hou (Fujitsu) > > wrote: > > > > > > Hi, > > > > > > When analyzing some issues in another thread[1], I found a bug that > > > fast fo

RE: Fix premature xmin advancement during fast forward decoding

2025-04-24 Thread Zhijie Hou (Fujitsu)
On Fri, Apr 25, 2025 at 5:44 AM Masahiko Sawada wrote: (Resending this email after compressing the attachment) > On Tue, Apr 22, 2025 at 12:06 AM Zhijie Hou (Fujitsu) > wrote: > > > > Hi, > > > > When analyzing some issues in another thread[1], I found a bug that > > fast forward decoding could

Re: Fix premature xmin advancement during fast forward decoding

2025-04-24 Thread Masahiko Sawada
Hi, On Tue, Apr 22, 2025 at 12:06 AM Zhijie Hou (Fujitsu) wrote: > > Hi, > > When analyzing some issues in another thread[1], I found a bug that fast > forward > decoding could lead to premature advancement of catalog_xmin, resulting in > required catalog data being removed by vacuum. > > The is

Re: Fix premature xmin advancement during fast forward decoding

2025-04-24 Thread Amit Kapila
On Tue, Apr 22, 2025 at 12:36 PM Zhijie Hou (Fujitsu) wrote: > > When analyzing some issues in another thread[1], I found a bug that fast > forward > decoding could lead to premature advancement of catalog_xmin, resulting in > required catalog data being removed by vacuum. > > The issue arises be

RE: Fix premature xmin advancement during fast forward decoding

2025-04-23 Thread Zhijie Hou (Fujitsu)
On Wed, Apr 23, 2025 at 2:31 PM shveta malik wrote: > > On Tue, Apr 22, 2025 at 12:36 PM Zhijie Hou (Fujitsu) > wrote: > > > > Hi, > > > > To fix this, I think we can allow the base snapshot to be built during fast > > forward decoding, as implemented in the patch 0001 (We already built base > >

Re: Fix premature xmin advancement during fast forward decoding

2025-04-22 Thread shveta malik
On Tue, Apr 22, 2025 at 12:36 PM Zhijie Hou (Fujitsu) wrote: > > Hi, > > To fix this, I think we can allow the base snapshot to be built during fast > forward decoding, as implemented in the patch 0001 (We already built base > snapshot in fast-forward mode for logical message in logicalmsg_decode(

Fix premature xmin advancement during fast forward decoding

2025-04-22 Thread Zhijie Hou (Fujitsu)
Hi, When analyzing some issues in another thread[1], I found a bug that fast forward decoding could lead to premature advancement of catalog_xmin, resulting in required catalog data being removed by vacuum. The issue arises because we do not build a base snapshot when decoding changes during fast