Re: increasing the default WAL segment size

Hi,

2)Getting "Aborted (core dumped)" error at the time of running
pg_basebackup , *(this issue is only coming on Linux32 ,not on Linux64)*
we have double check to confirm it .

Steps to reproduce on Linux32

> ===================
>> fetch the sources
>> apply both the patches
>> ./configure --with-zlib --enable-debug --enable-cassert
>> --enable-depend --prefix=$PWD/edbpsql --with-openssl CFLAGS="-g -O0"; make
>> all install
>> Performed initdb with switch "--wal-segsize 4"
>>
>
> Does the crash occur with only size 4?
>

Crash occurs for the value of *"--wal-segsize " 1, 2, 4, 8* with stack
details as below :

[bin]$ ./pg_basebackup -v -D /tmp/slave8
*** glibc detected *** ./pg_basebackup: free(): invalid pointer: 0x09fe9f00
***
======= Backtrace: =========
/lib/libc.so.6[0xbafb81]
/home/edb/PGsrc13march/postgresql/edbpsql/lib/libpq.so.5(PQclear+0x16d)[0x5696f5]
./pg_basebackup[0x8051441]
./pg_basebackup[0x804e7b5]
/lib/libc.so.6(__libc_start_main+0xe6)[0xb55d36]
./pg_basebackup[0x804a231]
======= Memory map: ========
00165000-001c7000 r-xp 00000000 08:03 1333807 /usr/lib/libssl.so.1.0.1e
001c7000-001ca000 r--p 00061000 08:03 1333807 /usr/lib/libssl.so.1.0.1e
001ca000-001ce000 rw-p 00064000 08:03 1333807 /usr/lib/libssl.so.1.0.1e
001ce000-0020c000 r-xp 00000000 08:03 1717206 /lib/libgssapi_krb5.so.2.2
0020c000-0020d000 r--p 0003e000 08:03 1717206 /lib/libgssapi_krb5.so.2.2
0020d000-0020e000 rw-p 0003f000 08:03 1717206 /lib/libgssapi_krb5.so.2.2
0020e000-002e5000 r-xp 00000000 08:03 1717208 /lib/libkrb5.so.3.3
002e5000-002eb000 r--p 000d6000 08:03 1717208 /lib/libkrb5.so.3.3
002eb000-002ec000 rw-p 000dc000 08:03 1717208 /lib/libkrb5.so.3.3
002ec000-00309000 r-xp 00000000 08:03 1706348
/lib/libgcc_s-4.4.7-20120601.so.1
00309000-0030a000 rw-p 0001d000 08:03 1706348
/lib/libgcc_s-4.4.7-20120601.so.1
00362000-00510000 r-xp 00000000 08:03 1333806
/usr/lib/libcrypto.so.1.0.1e
00510000-00520000 r--p 001ae000 08:03 1333806
/usr/lib/libcrypto.so.1.0.1e
00520000-00527000 rw-p 001be000 08:03 1333806
/usr/lib/libcrypto.so.1.0.1e
00527000-0052a000 rw-p 00000000 00:00 0
0055a000-00585000 r-xp 00000000 08:03 419296
/home/edb/PGsrc13march/postgresql/edbpsql/lib/libpq.so.5.10
00585000-00587000 rw-p 0002a000 08:03 419296
/home/edb/PGsrc13march/postgresql/edbpsql/lib/libpq.so.5.10
0086b000-0086e000 r-xp 00000000 08:03 1717205 /lib/libcom_err.so.2.1
0086e000-0086f000 r--p 00002000 08:03 1717205 /lib/libcom_err.so.2.1
0086f000-00870000 rw-p 00003000 08:03 1717205 /lib/libcom_err.so.2.1
008e3000-00900000 r-xp 00000000 08:03 1725674 /lib/libselinux.so.1
00900000-00901000 r--p 0001d000 08:03 1725674 /lib/libselinux.so.1
00901000-00902000 rw-p 0001e000 08:03 1725674 /lib/libselinux.so.1
00a10000-00a11000 r-xp 00000000 00:00 0 [vdso]
00af9000-00b03000 r-xp 00000000 08:03 1717209 /lib/libkrb5support.so.0.1
00b03000-00b04000 r--p 00009000 08:03 1717209 /lib/libkrb5support.so.0.1
00b04000-00b05000 rw-p 0000a000 08:03 1717209 /lib/libkrb5support.so.0.1
00b19000-00b37000 r-xp 00000000 08:03 1704925 /lib/ld-2.12.so
00b37000-00b38000 r--p 0001d000 08:03 1704925 /lib/ld-2.12.so
00b38000-00b39000 rw-p 0001e000 08:03 1704925 /lib/ld-2.12.so
00b3f000-00ccf000 r-xp 00000000 08:03 1704931 /lib/libc-2.12.so
00ccf000-00cd0000 ---p 00190000 08:03 1704931 /lib/libc-2.12.so
00cd0000-00cd2000 r--p 00190000 08:03 1704931 /lib/libc-2.12.so
00cd2000-00cd3000 rw-p 00192000 08:03 1704931 /lib/libc-2.12.so
00cd3000-00cd6000 rw-p 00000000 00:00 0
00cd8000-00cef000 r-xp 00000000 08:03 1704933 /lib/libpthread-2.12.so
00cef000-00cf0000 r--p 00016000 08:03 1704933 /lib/libpthread-2.12.so
00cf0000-00cf1000 rw-p 00017000 08:03 1704933 /lib/libpthread-2.12.so
00cf1000-00cf3000 rw-p 00000000 00:00 0
00cf5000-00cf8000 r-xp 00000000 08:03 1704977 /lib/libdl-2.12.so
00cf8000-00cf9000 r--p 00002000 08:03 1704977 /lib/libdl-2.12.so
00cf9000-00cfa000 rw-p 00003000 08:03 1704977 /lib/libdl-2.12.so
00d33000-00d45000 r-xp 00000000 08:03 1704980 /lib/libz.so.1.2.3
00d45000-00d46000 r--p 00011000 08:03 1704980 /lib/libz.so.1.2.3
00d46000-00d47000 rw-p 00012000 08:03 1704980 /lib/libz.so.1.2.3
00de4000-00e0c000 r-xp 00000000 08:03 1710117 /lib/libk5crypto.so.3.1
00e0c000-00e0d000 r--p 00028000 08:03 1710117 /lib/libk5crypto.so.3.1
00e0d000-00e0e000 rw-p 00029000 08:03 1710117 /lib/libk5crypto.so.3.1
00e0e000-00e0f000 rw-p 00000000 00:00 0
00e18000-00e1a000 r-xp 00000000 08:03 1710112 /lib/libkeyutils.so.1.3
00e1a000-00e1b000 r--p 00001000 08:03 1710112 /lib/libkeyutils.so.1.3
00e1b000-00e1c000 rw-p 00002000 08:03 1710112 /lib/libkeyutils.so.1.3
00f04000-00f10000 r-xp 00000000 08:03 1704932 /lib/libnss_files-2.12.so
00f10000-00f11000 r--p 0000b000 08:03 1704932 /lib/libnss_files-2.12.so
00f11000-00f12000 rw-p 0000c000 08:03 1704932 /lib/libnss_files-2.12.so
07be1000-07bf6000 r-xp 00000000 08:03 1704916 /lib/libresolv-2.12.so
07bf6000-07bf7000 ---p 00015000 08:03 1704916 /lib/libresolv-2.12.so
07bf7000-07bf8000 r--p 00015000 08:03 1704916 /lib/libresolv-2.12.so
07bf8000-07bf9000 rw-p 00016000 08:03 1704916 /lib/libresolv-2.12.so
07bf9000-07bfb000 rw-p 00000000 00:00 0
08048000-0805f000 r-xp 00000000 08:03 539967
/home/edb/PGsrc13march/postgresql/edbpsql/bin/pg_basebackup
0805f000-08060000 rw-p 00016000 08:03 539967
/home/edb/PGsrc13march/postgresql/edbpsql/bin/pg_basebackup
08060000-08062000 rw-p 00000000 00:00 0
09fe1000-0a002000 rw-p 00000000 00:00 0 [heap]
b74ec000-b76ec000 r--p 00000000 08:03 1333666
/usr/lib/locale/locale-archive
b76ec000-b76f1000 rw-p 00000000 00:00 0
b7700000-b7702000 rw-p 00000000 00:00 0
bfcf0000-bfd05000 rw-p 00000000 00:00 0 [stack]
Aborted (core dumped)

For value the value of *"--wal-segsize " 16, 32, 64*... (all multiple of
16) we are getting "Segmentation fault" message as below:
[bin]$ ./pg_basebackup -v -D /tmp/slave16
Segmentation fault (core dumped)

and for all other values of* "--wal-segsize " 3, 5, 7, 9, 10, 11, ... 15,
17, 18,* ... we are getting invalid message during "initdb":
[bin]$ ./initdb -D data1 --wal-segsize=17
initdb: Invalid WAL segment size 17

>
>
>> start the server
>> run pg_basebackup
>>
>> [centos(at)tushar-centos bin]$ ./pg_basebackup -v -D /tmp/myslave
>> *** glibc detected *** ./pg_basebackup: free(): invalid pointer:
>> 0x08da7f00 ***
>>
>> [centos(at)tushar-centos bin]$
>>
>> same scenario is working fine against HEAD (v10 ) on Linux32 [i.e no
>> patch applied]
>>
>> [centos(at)tushar-centos bin]$ ./pg_basebackup --verbose -D /tmp/slave11
>> pg_basebackup: initiating base backup, waiting for checkpoint to complete
>> pg_basebackup: checkpoint completed
>> pg_basebackup: transaction log start point: 0/2800024 on timeline 1
>> pg_basebackup: starting background WAL receiver
>> pg_basebackup: transaction log end point: 0/28000E4
>> pg_basebackup: waiting for background process to finish streaming ...
>> pg_basebackup: base backup completed
>> [centos(at)tushar-centos bin]$
>>
>
> Just to confirm, was this done with configure flag --with-wal-segsize=4 ?
>

we also have configure with the option "*--with-wal-segsize=4*" and getting
warning.
./configure --with-zlib --enable-debug --enable-cassert --enable-depend
--prefix=$PWD/inst --with-openssl CFLAGS="-g -O0" *--with-wal-segsize=4*

configure: WARNING: unrecognized options: --with-wal-segsize

Thanks & Regards,

Prabhat Kumar Sahu
EnterpriseDB: http://www.enterprisedb.com