Can't scan into dest[0]: cannot assign NULL to *float64


I’ve initialized a workload, which took around 10 hours:

cockroach workload init tpcc --warehouses=13000 'postgresql://root@wh13000crdb-cockroachdb-public:26257?sslmode=disable'

Then I ran it with:

cockroach workload run tpcc --warehouses=13000 --ramp=3m --duration=10m 'postgresql://root@wh13000crdb-cockroachdb-public:26257?sslmode=disable'

But to my dismay it throws this error:

Is there any workaround for this?
I assume it is because the initializing phase didn’t finish at it should. Is there any way to resume it without having to restart it from the beginning?

Here is how it looks in the console:


What version of cockroach are you running? What cluster topology?

I wouldn’t recommend running 13,000 warehouses in a single cockroach node. If you are running in a cluster topology, you should be pointing the workload binary to all of the nodes in your cluster.

You will need a high number of file descriptors available for your processes for this test, which will need 26,000 file descriptors on each process. I would recommend setting unlimited file descriptors for both processes using ulimit and trying again.


I think that is what I am doing. The image I am using is cockroach:v20.2.4

I deployed it like this:

helm install wh13000crdb cockroachdb/cockroachdb --namespace=thesis-crdb

Then I scaled it out to 15 nodes:

kubectl scale statefulset wh13000crdb-cockroachdb --replicas=15 --namespace=thesis-crdb

Tried it again but it stopped after 2 hours this time.

Giving it one last attempt with this:

I am pointing it to the wh13000crdb-cockroachdb-public.
First however I need to get inside one of the pods in order to execute cockroach commands:


It failed again, now after just 1 hour. I am out of ideas how to be able to get this workload initialized :frowning: