Thread: BUG #18612: Postgres crash with segfault on disk full - ____strtof_l_internal (strtod_l.c:1019)

The following bug has been logged on the website:

Bug reference:      18612
Logged by:          Bhupendra Patel
Email address:      bhavin.ec50@gmail.com
PostgreSQL version: 16.2
Operating system:   Linux
Description:

I have observed Postgres coredump upon disk full. Below is the stack trace
of core dump.

libc.so.6.  +0x0355e0 ____strtof_l_internal (strtod_l.c:1019)

libc.so.6.  +0x078fa0  mcheck_pedantic (mcheck.c:408)

libc.so.6.  +0x07780f __libc_realloc (malloc.c:3223)

postgres  +0x6e90bf _fini 

postgres  +0x58e9f8 errfinish 

postgres  +0x3dd7a1  CheckPointReplicationOrigin 

postgres +0x191e73  xact_redo 

postgres +0x198432 CreateCheckPoint 

postgres +0x0f92d9  _init 

postgres +0x3bc361  CheckpointerMain 

postgres +0x3ba6a5  AuxiliaryProcessMain 

postgres +0x3c0054  ClosePostmasterPorts 

postgres +0x3c37fb  PostmasterMain 

postgres +0x0faed9  main 

libc.so.6  +0x021a96 __gconv_lookup_cache (gconv_cache.c:367)

postgres +0x0faa4f  errstart_cold 

postgres. +0x0faf6f main

I have another version of Postgres 9.2 and haven't observed same on it.
Please take a look and let me know if require more details.

> On 12 Sep 2024, at 09:19, PG Bug reporting form <noreply@postgresql.org> wrote:

> I have observed Postgres coredump upon disk full. Below is the stack trace
> of core dump.

> I have another version of Postgres 9.2 and haven't observed same on it.
> Please take a look and let me know if require more details.

PostgreSQL doesn't cope all that well when disk is full, and I can imagine it
crashing at a number of different places even with the same version (let alone
one which shipped 12 years ago).

--
Daniel Gustafsson