diff options
| author | Donald Sharp <sharpd@nvidia.com> | 2021-02-02 12:56:06 -0500 |
|---|---|---|
| committer | Donald Sharp <sharpd@nvidia.com> | 2021-03-26 11:41:57 -0400 |
| commit | 039d547f6f1b7fe4b96cc22ac5a6ef8d18d5cf97 (patch) | |
| tree | f5614f80d048b8a3c4906f06284c45ef6373ec60 /lib/lib_errors.c | |
| parent | 694df37daf95940e923905fdd1096541d860bbfd (diff) | |
lib: Differentiate between real and cpu bound processes
When generating SLOW_THREAD warnings let's differentiate between
a cpu bound process and a wall bound process. Effectively
a slow thread can now be a process in FRR doing lots of work( cpu bound )
or wall bound ( the cpu is heavy load and a FRR process may be pre-empted
and never scheduled ).
Signed-off-by: Donald Sharp <sharpd@nvidia.com>
Diffstat (limited to 'lib/lib_errors.c')
| -rw-r--r-- | lib/lib_errors.c | 12 |
1 files changed, 9 insertions, 3 deletions
diff --git a/lib/lib_errors.c b/lib/lib_errors.c index 6e5088142a..17695e6607 100644 --- a/lib/lib_errors.c +++ b/lib/lib_errors.c @@ -45,9 +45,15 @@ static struct log_ref ferr_lib_warn[] = { .suggestion = "Gather log data and open an Issue. restart FRR", }, { - .code = EC_LIB_SLOW_THREAD, - .title = "The Event subsystem has detected a slow process", - .description = "The Event subsystem has detected a slow process, this typically indicates that FRR is having trouble completing work in a timely manner. This can be either a misconfiguration, bug, or some combination therof.", + .code = EC_LIB_SLOW_THREAD_CPU, + .title = "The Event subsystem has detected a slow cpu time process", + .description = "The Event subsystem has detected a slow process, this typically indicates that FRR is having trouble completing work in a timely manner. This can be either a misconfiguration, bug, or some combination therof. In this case total CPU time was over 5 seconds. Which indicates that FRR is very busy doing some work and should be addressed", + .suggestion = "Gather log data and open an Issue", + }, + { + .code = EC_LIB_SLOW_THREAD_WALL, + .title = "The Event subsystem has detected a slow wall time process", + .description = "The Event subsystem has detected a slow process, this typically indicates that FRR is having trouble completing work in a timely manner. This can be either a misconfiguration, bug or some combination therof. In this case total WALL time was over 5 seconds. Which indicates that FRR might be having trouble being scheduled or some system call is delaying", .suggestion = "Gather log data and open an Issue", }, { |