On 7 Oct 2025, at 16:43, Ilya Maximets wrote:
> These tests are executing group updates and then check the logs, but
> it is possible that logs are not yet flushed to the disk. That is
> causing sporadic failures like this one:
>
> +awk: stddev.awk:10: fatal: division by zero attempted
>
> This happens when we run the awk script to check the bucket load
> before the logs actually appear in the file.
>
> Fix that by always waiting for the logs.
>
> FreeBSD also has a long standing issue with async IO where some logs
> can be re-ordered and jammed together. This mostly happens in the
> 'equal weights' test due to high amount of logging. Changing the
> pattern from 'ofproto_dpif|DBG|.*Bucket' to a shorter '|DBG|.*Bucket'
> should solve most of the issues of this kind in this particular test.
>
> Signed-off-by: Ilya Maximets <[email protected]>
> ---
Thanks for looking into this. The changes look good to me.
Acked-by: Eelco Chaudron <[email protected]>
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
