Peter,
No we just noticed that we were unable to update that 1 record, and only that Boolean field. We did not search for
otherrecords that might have been affected.
I did as yet not run VACUUM, I’ll look into that later today.
Herman
Verstuurd vanaf mijn iPad
> Op 27 okt. 2021 om 03:45 heeft Peter Geoghegan <pg@bowt.ie> het volgende geschreven:
>
> On Mon, Oct 25, 2021 at 10:35 PM Herman verschooten
> <Herman@verschooten.net> wrote:
>> I can confirm that after dropping and recreating the 6 indexes everything is working fine.
>
> Did you also run VACUUM?
>
> You said that not all rows seem to be affected. Is there any pattern
> that you notice about the rows that are affected?
>
> I understand (from your remarks onSlack) that you found that the
> problem seemed to go away once you dropped the index named
> index_freights_on_cmr_received. But once you recreated the index from
> scratch, the problem came back.
>
> In other words, rebuilding the index didn't seem to make the "posting
> list tuple with 20 items cannot be split at offset 168" error go away,
> but dropping the index appeared to *temporarily* resolve the issue?
> Suggesting that the problem was either in the table, or in the code
> that builds a new index structure by scanning the table?
>
>> Thanks for all the help, both here and on slack,
>
> Thanks for the report!
>
> --
> Peter Geoghegan