Hi Thang,
Good catch !
Can you please let me know if you were starting pm monitoring on a component
pid, which is in root or non-root?
The issue, you are reporting, Thang, may come when Amfnd' (kill) doesn't have
permission to send signal to the pid.
I am assuming that you are getting EPERM as a return.
Can you please verify, Thang(by logging, just like done in mon.cc at "switch
(kill(mon_rec->pid, 0)) {"), if the return is EPERM or something else?
If it is correct, then ideally, the user shouldn't be using pm monitoring on
such pid and that means no change is needed in the patch.
can you please confirm, Thang.
Thanks
Anand Sundararaj
Senior Solutions Architect | 480 686 4772
www.GetHighAvailability.com
Get High Availability Today!
NJ, USA: +1 508-507-6507
> On 07/20/2020 9:11 PM Thang Duc Nguyen <thang.d.ngu...@dektech.com.au> wrote:
>
>
> I missed one info. I tested with NON_ROOT user.
> export AMFND_NON_ROOT=1
>
> -----Original Message-----
> From: Thang Duc Nguyen <thang.d.ngu...@dektech.com.au>
> Sent: Tuesday, July 21, 2020 11:02 AM
> To: s.an...@gethighavailability.com; Minh Hon Chau
> <minh.c...@dektech.com.au>; nagen...@gethighavailability.com;
> p...@gethighavailability.com
> Cc: opensaf-devel@lists.sourceforge.net
> Subject: Re: [devel] [PATCH 1/1] amf: remove hard-coding in amfnd [#3205]
>
> Hi Sundararaj,
>
> The Opensaf can start but it may not run correctly.
> I tested your patch. During passive mornitoring process of copoenent, this
> patch cuases the coredump in function avnd_comp_pm_param_val() when invoking
> saAmfPmStop()/saAmfPmStart().
>
> Snipest code indicate the crash in cpm.cc file void avnd_comp_pm_param_val( )
> { ...
> if (kill(pm_start->pid, 0) == -1) {
> osafassert(errno == ESRCH); //Crash here due to retured error:
> Operation not permitted
> *o_amf_rc = SA_AIS_ERR_NOT_EXIST;
> return;
> }
> }
>
> B.R/Thang
> -----Original Message-----
> From: s.an...@gethighavailability.com <s.an...@gethighavailability.com>
> Sent: Tuesday, July 21, 2020 9:22 AM
> To: Minh Hon Chau <minh.c...@dektech.com.au>; Thang Duc Nguyen
> <thang.d.ngu...@dektech.com.au>; nagen...@gethighavailability.com;
> p...@gethighavailability.com
> Cc: opensaf-devel@lists.sourceforge.net; Anand Sundararaj
> <s.an...@gethighavailability.com>
> Subject: [PATCH 1/1] amf: remove hard-coding in amfnd [#3205]
>
> From: Anand Sundararaj <s.an...@gethighavailability.com>
>
> ---
> src/amf/amfnd/amfnd.conf | 3 +++
> src/amf/amfnd/main.cc | 9 ++++++++-
> 2 files changed, 11 insertions(+), 1 deletion(-)
>
> diff --git a/src/amf/amfnd/amfnd.conf b/src/amf/amfnd/amfnd.conf index
> 07bc0ba..4e8b07a 100644
> --- a/src/amf/amfnd/amfnd.conf
> +++ b/src/amf/amfnd/amfnd.conf
> @@ -34,3 +34,6 @@ export AVND_PM_MONITORING_RATE=1000 # It can be disabled
> if set THREAD_TRACE_BUFFER as 0, the maximum value # can be set as 65535.
> # export THREAD_TRACE_BUFFER=10240
> +
> +#AMFND run as root. Uncomment next line to run as a user mentioned in
> nid.conf.
> +#export AMFND_NON_ROOT=1
> diff --git a/src/amf/amfnd/main.cc b/src/amf/amfnd/main.cc index
> d7857fa..6d9ee95 100644
> --- a/src/amf/amfnd/main.cc
> +++ b/src/amf/amfnd/main.cc
> @@ -164,6 +164,7 @@ static void new_handler() {
>
> int main(int argc, char *argv[]) {
> uint32_t error;
> + char *val;
>
> // function to be called if new fails. The alternative of using catch of
> // std::bad_alloc will unwind the stack and thus no call chain will be @@
> -179,7 +180,13 @@ int main(int argc, char *argv[]) {
> goto done;
> }
>
> - daemonize_as_user("root", argc, argv);
> + if ((val = getenv("AMFND_NON_ROOT")) != nullptr) {
> + daemonize(argc, argv);
> + TRACE("AMFND will run as non-root"); } else {
> + daemonize_as_user("root", argc, argv);
> + TRACE("AMFND will run as root");
> + }
>
> // Enable long DN
> if (setenv("SA_ENABLE_EXTENDED_NAMES", "1", 1) != 0) {
> --
> 2.7.4
>
>
>
> _______________________________________________
> Opensaf-devel mailing list
> Opensaf-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/opensaf-devel
_______________________________________________
Opensaf-devel mailing list
Opensaf-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opensaf-devel
