Increase timeout for nptl/tst-robust-fork
Commit Message
Although tst-robust-fork usually completes in less the 20s, there are
some slower machines which may take as long as 23s.
This patch increases the timeout to 35s in order to minimize
intermittent failures on slower machines and under heavy usage.
2017-02-08 Tulio Magno Quites Machado Filho <tuliom@linux.vnet.ibm.com>
* nptl/tst-robust-fork.c (TIMEOUT): Set to 35.
---
nptl/tst-robust-fork.c | 2 ++
1 file changed, 2 insertions(+)
Comments
On 02/08/2017 02:17 PM, Tulio Magno Quites Machado Filho wrote:
> Although tst-robust-fork usually completes in less the 20s, there are
> some slower machines which may take as long as 23s.
> This patch increases the timeout to 35s in order to minimize
> intermittent failures on slower machines and under heavy usage.
Can't you increase TIMEOUTFACTOR on such machines?
Nowadays, it should not scale expected timeouts anymore.
Thanks,
Florian
Florian Weimer <fweimer@redhat.com> writes:
> On 02/08/2017 02:17 PM, Tulio Magno Quites Machado Filho wrote:
>> Although tst-robust-fork usually completes in less the 20s, there are
>> some slower machines which may take as long as 23s.
>> This patch increases the timeout to 35s in order to minimize
>> intermittent failures on slower machines and under heavy usage.
>
> Can't you increase TIMEOUTFACTOR on such machines?
Yes, I can, but this is the only testcase timing out on those machines.
Adjusting the timeout of this testcase will prevent us from having
extra steps or settings to the way we test glibc instead of the usual
configure && make && make check.
Why can't we adjust the timeout of this testcase to avoid those spurious
failures?
@@ -25,6 +25,8 @@
#include <support/xunistd.h>
#include <sys/mman.h>
+#define TIMEOUT 35
+
/* Data shared between processes. */
struct shared
{