Re: BUG #19930: amcheck might report btree index corruption after recovery of a mid-deletion state

From: Alexander Lakhin <exclusion(at)gmail(dot)com>
To: pgsql-bugs(at)lists(dot)postgresql(dot)org
Subject: Re: BUG #19930: amcheck might report btree index corruption after recovery of a mid-deletion state
Date: 2026-11-22 18:00:00
Message-ID: b45631ed-aaa3-44aa-9d8a-2d2acdf1ee51@gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

22.11.2026 19:00, PG Bug reporting form wrote:
> The following bug has been logged on the website:
>
> Bug reference: 19930
> ...
>
> In case a multi-level btree deletion is interrupted by crash, amcheck can
> report XX002 error:
> ERROR: the first child of leftmost target page is not leftmost of its level
> in index "xxx"
> DETAIL: Target block=xxx child block=xxx target page lsn=X/X.
> when executed right after recovery (before VACUUM, etc).
>
> Commit 6ec9e9975 fixed a similar issue but for single-level deletions only.
>
> A demo test to be attached.

Please find the demo test attached. It works/fails for me as below:
make -s check -C contrib/amcheck/
...
#   Failed test 'bt_index_parent_check passes at the interrupted deletion'
#   at t/099_pitr_multilevel.pl line 87.
#          got: '3'
#     expected: '0'

#   Failed test 'no XX002 leftmost finding'
#   at t/099_pitr_multilevel.pl line 88.
#                   'psql:<stdin>:1: ERROR:  the first child of leftmost target page is not leftmost of its level in
index "no_index_cleanup_idx"
# DETAIL:  Target block=8 child block=7 target page lsn=0/0311A9C8.'
#           matches '(?^:first child of leftmost target page is not leftmost of its level)'
# Looks like you failed 2 tests of 4.
t/099_pitr_multilevel.pl ....... Dubious, test returned 2 (wstat 512, 0x200)
Failed 2/4 subtests

Best regards,
Alexander

Attachment Content-Type Size
099_pitr_multilevel.pl application/x-perl 3.9 KB

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Raja Sai pranav 2026-11-22 19:10:50 Re: BUG #19930: amcheck might report btree index corruption after recovery of a mid-deletion state
Previous Message Nitin Motiani 2026-11-22 17:31:32 Re: BUG #19761: btree_gist: "cannot subtract infinite dates" when inserting 'infinity' into a GiST index on date