Hi,
On Thu, Aug 18, 2022 at 10:34 PM Justin Pryzby <pry...@telsasoft.com> wrote:
>
> Immediately after upgrading an internal instance, a loop around "vacuum" did
> this:
Thank you for the report!
>
> TRAP: FailedAssertion("indstats->status == PARALLEL_INDVAC_STATUS_INITIAL",
> File: "vacuumparallel.c", Line: 611, PID: 27635)
> postgres: postgres pryzbyj [local] VACUUM(ExceptionalCondition+0x8d)[0x99d9fd]
> postgres: postgres pryzbyj [local] VACUUM[0x6915db]
> postgres: postgres pryzbyj [local] VACUUM(heap_vacuum_rel+0x12b6)[0x5083e6]
> postgres: postgres pryzbyj [local] VACUUM[0x68e97a]
> postgres: postgres pryzbyj [local] VACUUM(vacuum+0x48e)[0x68fe9e]
> postgres: postgres pryzbyj [local] VACUUM(ExecVacuum+0x2ae)[0x69065e]
> postgres: postgres pryzbyj [local]
> VACUUM(standard_ProcessUtility+0x530)[0x8567b0]
> /usr/pgsql-15/lib/pg_stat_statements.so(+0x5450)[0x7f52b891c450]
> postgres: postgres pryzbyj [local] VACUUM[0x85490a]
> postgres: postgres pryzbyj [local] VACUUM[0x854a53]
> postgres: postgres pryzbyj [local] VACUUM(PortalRun+0x179)[0x855029]
> postgres: postgres pryzbyj [local] VACUUM[0x85099b]
> postgres: postgres pryzbyj [local] VACUUM(PostgresMain+0x199a)[0x85268a]
> postgres: postgres pryzbyj [local] VACUUM[0x496a21]
> postgres: postgres pryzbyj [local] VACUUM(PostmasterMain+0x11c0)[0x7b3980]
> postgres: postgres pryzbyj [local] VACUUM(main+0x1c6)[0x4986a6]
> /lib64/libc.so.6(__libc_start_main+0xf5)[0x7f52c4b893d5]
> postgres: postgres pryzbyj [local] VACUUM[0x498c59]
> < 2022-08-18 07:56:51.963 CDT >LOG: server process (PID 27635) was
> terminated by signal 6: Aborted
> < 2022-08-18 07:56:51.963 CDT >DETAIL: Failed process was running: VACUUM
> ANALYZE alarms
>
> Unfortunately, it looks like the RPM packages are compiled with -O2, so this
> is
> of limited use.
>
> Core was generated by `postgres: postgres pryzbyj [local] VACUUM
> '.
> Program terminated with signal 6, Aborted.
> #0 0x00007f52c4b9d207 in raise () from /lib64/libc.so.6
> Missing separate debuginfos, use: debuginfo-install
> audit-libs-2.8.4-4.el7.x86_64 bzip2-libs-1.0.6-13.el7.x86_64
> cyrus-sasl-lib-2.1.26-23.el7.x86_64 elfutils-libelf-0.176-5.el7.x86_64
> elfutils-libs-0.176-5.el7.x86_64 glibc-2.17-260.el7_6.3.x86_64
> keyutils-libs-1.5.8-3.el7.x86_64 krb5-libs-1.15.1-51.el7_9.x86_64
> libattr-2.4.46-13.el7.x86_64 libcap-2.22-9.el7.x86_64
> libcap-ng-0.7.5-4.el7.x86_64 libcom_err-1.42.9-19.el7.x86_64
> libgcc-4.8.5-39.el7.x86_64 libgcrypt-1.5.3-14.el7.x86_64
> libgpg-error-1.12-3.el7.x86_64 libicu-50.1.2-17.el7.x86_64
> libselinux-2.5-15.el7.x86_64 libstdc++-4.8.5-39.el7.x86_64
> libxml2-2.9.1-6.el7_9.6.x86_64 libzstd-1.5.2-1.el7.x86_64
> lz4-1.7.5-2.el7.x86_64 nspr-4.19.0-1.el7_5.x86_64 nss-3.36.0-7.1.el7_6.x86_64
> nss-softokn-freebl-3.36.0-5.el7_5.x86_64 nss-util-3.36.0-1.1.el7_6.x86_64
> openldap-2.4.44-21.el7_6.x86_64 openssl-libs-1.0.2k-22.el7_9.x86_64
> pam-1.1.8-22.el7.x86_64 pcre-8.32-17.el7.x86_64
> systemd-libs-219-62.el7_6.5.x86_64 xz-libs-5.2.2-1.el7.x86_64
> zlib-1.2.7-18.el7.x86_64
> (gdb) bt
> #0 0x00007f52c4b9d207 in raise () from /lib64/libc.so.6
> #1 0x00007f52c4b9e8f8 in abort () from /lib64/libc.so.6
> #2 0x000000000099da1e in ExceptionalCondition
> (conditionName=conditionName@entry=0xafae40 "indstats->status ==
> PARALLEL_INDVAC_STATUS_INITIAL", errorType=errorType@entry=0x9fb4b7
> "FailedAssertion",
> fileName=fileName@entry=0xafb0c0 "vacuumparallel.c",
> lineNumber=lineNumber@entry=611) at assert.c:69
> #3 0x00000000006915db in parallel_vacuum_process_all_indexes (pvs=0x2e85f80,
> num_index_scans=<optimized out>, vacuum=<optimized out>) at
> vacuumparallel.c:611
> #4 0x00000000005083e6 in heap_vacuum_rel (rel=<optimized out>,
> params=<optimized out>, bstrategy=<optimized out>) at vacuumlazy.c:2679
It seems that parallel_vacuum_cleanup_all_indexes() got called[1],
which means this was the first time to perform parallel vacuum (i.e.,
index cleanup).
I'm not convinced yet but it could be a culprit that we missed doing
memset(0) for the shared array of PVIndStats in
parallel_vacuum_init(). This shared array was introduced in PG15.
[1]
https://github.com/postgres/postgres/blob/REL_15_STABLE/src/backend/access/heap/vacuumlazy.c#L2679
Regards,
--
Masahiko Sawada
EDB: https://www.enterprisedb.com/
