On 1/10/24 12:25, Eelco Chaudron wrote: > Currently the 'Spent an unreasonably long Xms dumping flows' message > is set to the INFO level. However, based on this, we are also > drastically limiting the number of flows in the datapath, and this > would warrant a WARNING level. > > Acked-by: Simon Horman <[email protected]> > Signed-off-by: Eelco Chaudron <[email protected]> > --- > ofproto/ofproto-dpif-upcall.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/ofproto/ofproto-dpif-upcall.c b/ofproto/ofproto-dpif-upcall.c > index cc10f57b5..cd71e3ee3 100644 > --- a/ofproto/ofproto-dpif-upcall.c > +++ b/ofproto/ofproto-dpif-upcall.c > @@ -1049,7 +1049,7 @@ udpif_revalidator(void *arg) > atomic_store_relaxed(&udpif->flow_limit, flow_limit); > > if (duration > 2000) { > - VLOG_INFO("Spent an unreasonably long %lldms dumping flows", > + VLOG_WARN("Spent an unreasonably long %lldms dumping flows", > duration); > } >
Hmm. Interestingly, this is causing random test failures when a large time warp is happening during revalidation. For example I saw this test failing in CI: drop-stats.at:132: testing drop-stats - stack too deep Because of: 2024-01-15T23:11:09.413Z|00001|ofproto_dpif_upcall(revalidator6)|WARN|Spent an unreasonably long 5000ms dumping flows We probably need to adjust the tests to avoid large time warping where possible or exclude the warning from checking. Best regards, Ilya Maximets. _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
