[Bugs] [Bug 1512691] PostgreSQL DB Restore: unexpected data beyond EOF

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1512691



--- Comment #22 from Vijay Bellur <vbellur at redhat.com> ---
While running tests on a build with patches 20413 and 20410, I observed the
following:

$pgbench -c 8 -j 2 -T 30 bench-db
starting vacuum...WARNING:  corrupted statistics file "pg_stat_tmp/pgstat.stat"
end.

This is usually seen in the second iteration of a pgbench loop. First iteration
ran clean quite a few times.

Disabling quick-read seems to eliminate this problem. Relevant parts from
attached strace output (comment 21):

586436 open("pg_stat_tmp/pgstat.stat", O_RDONLY) = 12
586436 fstat(12, {st_mode=S_IFREG|0600, st_size=13479, ...}) = 0
586436 mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0)
= 0x7f57c5dec000
586436 read(12,
"\232\274\245\1\353\203Q\361\316\22\2\0\2\0\0\0\0\0\0\0\245\0\0\0\0\0\0\0+l\"\0"...,
8192) = 8192
586436 close(12)                        = 0
586436 munmap(0x7f57c5dec000, 8192)     = 0
586436 sendto(10, "\1\0\0\0\20\0\0\0])W\362\316\22\2\0", 16, 0, NULL, 0) = 16
586436 select(0, NULL, NULL, NULL, {0, 10000}) = 0 (Timeout)
586436 open("pg_stat_tmp/pgstat.stat", O_RDONLY) = 12
586436 fstat(12, 0x7fff84995980)        = -1 ESTALE (Stale file handle)
586436 mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0)
= 0x7f57c5dec000
586436 read(12, 0x7f57c5dec000, 8192)   = -1 ESTALE (Stale file handle)
586436 write(2, "\0\0>\0\304\362\10\0tWARNING:  corrupted sta"..., 71) = 71
586436 sendto(11, "N\0\0\0{SWARNING\0C01000\0Mcorrupted "..., 124, 0, NULL, 0)
= 124
586436 close(12)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug https://bugzilla.redhat.com/token.cgi?t=F4LvduhjjU&a=cc_unsubscribe