Thread: Modifying data type of slot_keep_segs from XLogRecPtr to XLogSegNo
Hi, Currently, slot_keep_segs is defined as "XLogRecPtr" in KeepLogSeg(), but it seems that should be "XLogSegNo" because this variable is segment number. How do you think? Regards, -- Atsushi Torikoshi NTT DATA CORPORATION
Attachment
On 2020/07/08 11:02, torikoshia wrote: > Hi, > > Currently, slot_keep_segs is defined as "XLogRecPtr" in KeepLogSeg(), > but it seems that should be "XLogSegNo" because this variable is > segment number. > > How do you think? I agree that using XLogRecPtr for slot_keep_segs is incorrect. But this variable indicates the number of segments rather than segment no, uint64 seems better. Thought? Regards, -- Fujii Masao Advanced Computing Technology Center Research and Development Headquarters NTT DATA CORPORATION
On 2020-07-08 11:15, Fujii Masao wrote: > On 2020/07/08 11:02, torikoshia wrote: >> Hi, >> >> Currently, slot_keep_segs is defined as "XLogRecPtr" in KeepLogSeg(), >> but it seems that should be "XLogSegNo" because this variable is >> segment number. >> >> How do you think? > > I agree that using XLogRecPtr for slot_keep_segs is incorrect. > But this variable indicates the number of segments rather than > segment no, uint64 seems better. Thought? That makes sense. The number of segments and segment number are different. Regards, -- Atsushi Torikoshi NTT DATA CORPORATION
On 2020/07/08 11:55, torikoshia wrote: > On 2020-07-08 11:15, Fujii Masao wrote: >> On 2020/07/08 11:02, torikoshia wrote: >>> Hi, >>> >>> Currently, slot_keep_segs is defined as "XLogRecPtr" in KeepLogSeg(), >>> but it seems that should be "XLogSegNo" because this variable is >>> segment number. >>> >>> How do you think? >> >> I agree that using XLogRecPtr for slot_keep_segs is incorrect. >> But this variable indicates the number of segments rather than >> segment no, uint64 seems better. Thought? > > That makes sense. > The number of segments and segment number are different. Yes, so patch attached. I will commit it later. Regards, -- Fujii Masao Advanced Computing Technology Center Research and Development Headquarters NTT DATA CORPORATION
Attachment
On 2020/07/08 15:22, Fujii Masao wrote: > > > On 2020/07/08 11:55, torikoshia wrote: >> On 2020-07-08 11:15, Fujii Masao wrote: >>> On 2020/07/08 11:02, torikoshia wrote: >>>> Hi, >>>> >>>> Currently, slot_keep_segs is defined as "XLogRecPtr" in KeepLogSeg(), >>>> but it seems that should be "XLogSegNo" because this variable is >>>> segment number. >>>> >>>> How do you think? >>> >>> I agree that using XLogRecPtr for slot_keep_segs is incorrect. >>> But this variable indicates the number of segments rather than >>> segment no, uint64 seems better. Thought? >> >> That makes sense. >> The number of segments and segment number are different. > > Yes, so patch attached. I will commit it later. Pushed. Thanks! Regards, -- Fujii Masao Advanced Computing Technology Center Research and Development Headquarters NTT DATA CORPORATION
At Wed, 8 Jul 2020 21:27:04 +0900, Fujii Masao <masao.fujii@oss.nttdata.com> wrote in > > > On 2020/07/08 15:22, Fujii Masao wrote: > > On 2020/07/08 11:55, torikoshia wrote: > >> On 2020-07-08 11:15, Fujii Masao wrote: > >>> On 2020/07/08 11:02, torikoshia wrote: > >>>> Hi, > >>>> > >>>> Currently, slot_keep_segs is defined as "XLogRecPtr" in KeepLogSeg(), > >>>> but it seems that should be "XLogSegNo" because this variable is > >>>> segment number. > >>>> > >>>> How do you think? Yeah, that's my mistake while made bouncing back and forth between segments and LSN in the code. I noticed that once but forgotten until now. Thanks for finding it. > >>> I agree that using XLogRecPtr for slot_keep_segs is incorrect. > >>> But this variable indicates the number of segments rather than > >>> segment no, uint64 seems better. Thought? > >> > >> That makes sense. > >> The number of segments and segment number are different. > > Yes, so patch attached. I will commit it later. > > Pushed. Thanks! Thanks! -- Kyotaro Horiguchi NTT Open Source Software Center