-
Problem report
-
Resolution: Duplicate
-
Trivial
-
None
-
5.4.2
proc.num returns the wrong value.
this happens for multiple proc.num checks in our environment.
service uptime = 4 months
**i believe this issue is related to this line of code: here
for entries, err = f.Readdir(1); err != io.EOF; entries, err = f.Readdir(1) {
working strace: contains the proc or the process being opened.
newfstatat(AT_FDCWD, "/proc/31745", {st_mode=S_IFDIR|0555, st_size=0, ...}, AT_SYMLINK_NOFOLLOW) = 0 8271 <... nanosleep resumed>NULL) = 0 8274 openat(AT_FDCWD, "/proc/31745/stat", O_RDONLY) = 5 8271 futex(0xc000478148, FUTEX_WAKE_PRIVATE, 1 <unfinished ...> 8274 read(5, <unfinished ...> 8286 <... futex resumed>) = 0 8274 <... read resumed>"31745 (sssd) S 1 31745 31745 0 -1 1077944576 1908 0 29 0 468 1189 0 0 20 0 1 0 1782782995 275206144 947 18446744073709551615 94783524737024 94783524813844 140724449739856 140724449738568 140195446729904 0 4224 1052672 84483 18446744072478930117 0 0 17 2 0 0 3 0 0 94783526912744 94783526916512 94783536095232 140724449742658 140724449742691 140724449742691 140724449742825 0\n", 2048) = 375 8271 <... futex resumed>) = 1 8286 futex(0xc000478148, FUTEX_WAIT_PRIVATE, 0, NULL <unfinished ...> 8274 close(5 <unfinished ...> 8286 <... futex resumed>) = -1 EAGAIN (Resource temporarily unavailable) 8274 <... close resumed>) = 0 8271 futex(0xc000478148, FUTEX_WAKE_PRIVATE, 1 <unfinished ...> 8286 futex(0xc000478148, FUTEX_WAIT_PRIVATE, 0, NULL <unfinished ...> 8274 openat(AT_FDCWD, "/proc/31745/cmdline", O_RDONLY) = 5
not working strace: last newfstatat systemcall looking at the /proc files
8441 newfstatat(AT_FDCWD, "/proc/8454", 0xc00038b628, AT_SYMLINK_NOFOLLOW) = -1 ENOENT (No such file or directory)
This server has a lot of processes running (>700)
I think between the time of assigning the variable here
f, err := os.Open("/proc")
and the for loop running the process (8441) ends causing the read error which stops the loop?
- duplicates
-
-
- Closed
-