On 2/13/19 7:28 AM, Pawel Baldysiak wrote:
> Since the patch
> c76242c5("mdmon: get safe mode delay file descriptor early"),
> safe_mode_dalay is set properly by initrd mdmon.
> But in some cases with filesystem traffic since the very
> start of the system, it might take a while to transit to clean state.
> Due to fact that new mdmon does not wait for the old one to exit -
> it might happen that the new one switches safe_mode_delay back to
> seconds, before old one exits. As the result two mdmons are running
> concurrently on same array.
>
> Wait for the old mdmon to exit by pinging it with SIGUSR1 signal,
> just in case it is sleeping.
>
> Signed-off-by: Pawel Baldysiak <pawel.baldysiak@xxxxxxxxx>
> ---
> mdmon.c | 14 +++++++++++---
> 1 file changed, 11 insertions(+), 3 deletions(-)
>
> diff --git a/mdmon.c b/mdmon.c
> index 0955fcc5..03e6e427 100644
> --- a/mdmon.c
> +++ b/mdmon.c
> @@ -171,6 +171,7 @@ static void try_kill_monitor(pid_t pid, char *devname, int sock)
> int fd;
> int n;
> long fl;
> + int rv;
>
> /* first rule of survival... don't off yourself */
> if (pid == getpid())
> @@ -201,9 +202,16 @@ static void try_kill_monitor(pid_t pid, char *devname, int sock)
> fl &= ~O_NONBLOCK;
> fcntl(sock, F_SETFL, fl);
> n = read(sock, buf, 100);
> - /* Ignore result, it is just the wait that
> - * matters
> - */
> +
> + /* If there is I/O going on it might took some time to get to
> + * clean state. Wait for monitor to exit fully to avoid races.
> + * Ping it with SIGUSR1 in case that it is sleeping */
> + do {
> + rv = kill(pid, SIGUSR1);
> + if (rv < 0)
> + break;
> + usleep(200000);
> + } while (1);
I think the principle of this is fine, but doing an indefinite while(1)
raises the little hairs on the back of my neck. Could you limit it to
say 5 or 10 runs?
Thanks,
Jes
