> On one of my servers stat collector went crasy on idle system.
> It constantly write around 50MB/s on idle database (no activity at all).
>
Detailed strace look like endless tight loop of:
... 5MB writes ...
21:03:01.131550 write(7,
"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
2188) = 2188
21:03:01.131613 close(7) = 0
21:03:01.131674 munmap(0x7fde72dc1000, 4096) = 0
21:03:01.131743 rename("/mnt/tmpfs/pgstat.tmp", "/mnt/tmpfs/pgstat.stat") =
0
21:03:01.131826 recvfrom(11, 0x7fff4148a620, 1000, 0, 0, 0) = -1 EAGAIN
(Resource temporarily unavailable)
21:03:01.131889 read(3, 0x7fff4148a5b0, 16) = -1 EAGAIN (Resource
temporarily unavailable)
21:03:01.131942 poll([{fd=11, events=POLLIN}, {fd=3, events=POLLIN}, {fd=8,
events=POLLIN}], 3, -1) = 1 ([{fd=11
, revents=POLLIN}])
21:03:01.144257 recvfrom(11,
"\10\0\0\0\30\0\0\0V\277\363\23\0\0\0\0(\262k\233\305\253\1\0", 1000, 0,
NULL, NULL
) = 24
21:03:01.144354 recvfrom(11, 0x7fff4148a620, 1000, 0, 0, 0) = -1 EAGAIN
(Resource temporarily unavailable)
21:03:01.144409 read(3, 0x7fff4148a5b0, 16) = -1 EAGAIN (Resource
temporarily unavailable)
21:03:01.144479 poll([{fd=11, events=POLLIN}, {fd=3, events=POLLIN}, {fd=8,
events=POLLIN}], 3, -1) = 1 ([{fd=11
, revents=POLLIN}])
21:03:01.146321 recvfrom(11, "\1\0\0\0\20\0\0\0X\223k\233\305\253\1\0",
1000, 0, NULL, NULL) = 16
21:03:01.146382 open("/mnt/tmpfs/pgstat.tmp", O_WRONLY|O_CREAT|O_TRUNC,
0666) = 7
21:03:01.146458 fstat(7, {st_mode=S_IFREG|0600, st_size=0, ...}) = 0
21:03:01.146519 mmap(NULL, 4096, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7fde72dc1000
21:03:01.146592 write(7,
"\232\274\245\1N\273k\233\305\253\1\0\374\27\1\0\0\0\0\0\250\7\0\0\0\0\0\0\311\331\243\
270"..., 4096) = 4096
21:03:01.146664 write(7,
"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
4096) = 4096
... 5MB writes...
and so on, with no visible changes...
Kind Regards,
Maksym
