cnf-tests: Add debug information for flake tests #2109
Conversation
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: zeeke The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
openshift-ci
bot
added
the
approved
| // [0200] is the class for Ethernet controllers | ||
| // https://admin.pci-ids.ucw.cz/read/PD/ | ||
| out, err := testnode.ExecCommandOnNodeViaSriovDaemon(client.Client, node, []string{ | ||
| "lspci", "-vv", "-nn", "-d", "::0200", |
There was a problem hiding this comment.
and so we introduce a hard dependency on both lspci and the pcidb. Is this a concern?
There was a problem hiding this comment.
good point. I didn't see it from that perspective.
What about using a command like:
sh-5.1# sh -c 'for d in /sys/class/net/*/device/numa_node; do echo -n "$d "; cat ${d}; done'
/sys/class/net/eno12399/device/numa_node 0
/sys/class/net/eno12409/device/numa_node 0
/sys/class/net/eno12409v0/device/numa_node 0
/sys/class/net/eno12409v1/device/numa_node 0
/sys/class/net/eno12409v3/device/numa_node 0
/sys/class/net/eno8303/device/numa_node 0
/sys/class/net/eno8403/device/numa_node 0
/sys/class/net/ens2f0np0/device/numa_node 1
/sys/class/net/ens2f1np1/device/numa_node 1
?
There was a problem hiding this comment.
that's better in the sense we don't add the implicit dep, and WORKSFORME
There was a problem hiding this comment.
As sidenote, we keep reinventing over and over custom solutiuons for the hw discovery/inventory problem, in both tests and application code
zeeke
force-pushed
the
numa-flake-cant-find-suitable-node
branch
2 times, most recently
from
31517f3
to
aa314d7
Compare
|
/test ? |
|
@zeeke: The following commands are available to trigger required jobs:
The following commands are available to trigger optional jobs:
Use
In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
|
/test e2e-telco5g-cnftests |
Test case ``` [sriov] NUMA node alignment [BeforeAll] Validate the creation of a pod with excludeTopology set to False and an SRIOV interface in a different NUMA node than the pod ``` flakes with the following error: ``` Can't find a suitable node for testing: node [cnfdu3] has no NUMA0 devices, node [cnfdu4] has no NUMA0 devices, ``` Dump NUMA nodes for each network device when failing. Use Gomega instead of `ginkgo.Fail` to trigger a k8sreporter archive Signed-off-by: Andrea Panattoni <[email protected]>
zeeke
force-pushed
the
numa-flake-cant-find-suitable-node
branch
from
aa314d7
to
2167379
Compare
|
/test e2e-telco5g-cnftests |
2 similar comments
|
/test e2e-telco5g-cnftests |
|
/test e2e-telco5g-cnftests |
|
These changes spotted a problem in the ExecCommand API: as the command returns some artifacts (e.g. |
Signed-off-by: Andrea Panattoni <[email protected]>
|
/test e2e-telco5g-cnftests |
2 similar comments
|
/test e2e-telco5g-cnftests |
|
/test e2e-telco5g-cnftests |
|
@zeeke: The following tests failed, say
Full PR test history. Your PR dashboard. Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here. |
Test case
flakes with the following error:
Dump NUMA nodes for each network device when failing.
Use Gomega instead of
ginkgo.Failto trigger a k8sreporter archive.Sample failure