helm, info, status: Check and report reachability sequencer + refactor#6073
Conversation
b376a39 to
b4302e1
Compare
96d24ab to
69a89d9
Compare
69a89d9 to
01a3ced
Compare
|
/cluster_test |
|
Deploy cluster test triggered for Commit 01a3ced14e736c8d92fc1b7b9fb3d072cb36585b in , please contact a Contributor to approve it in CircleCI: https://app.circleci.com/pipelines/github/DACH-NY/canton-network-internal/71127 |
|
/cluster_test |
|
Deploy cluster test triggered for Commit 01a3ced14e736c8d92fc1b7b9fb3d072cb36585b in , please contact a Contributor to approve it in CircleCI: https://app.circleci.com/pipelines/github/DACH-NY/canton-network-internal/null |
|
/cluster_test |
|
Deploy cluster test triggered for Commit 01a3ced14e736c8d92fc1b7b9fb3d072cb36585b in , please contact a Contributor to approve it in CircleCI: https://app.circleci.com/pipelines/github/DACH-NY/canton-network-internal/71265 |
|
martinflorian-da
left a comment
There was a problem hiding this comment.
I agree with Itai that this is hard to review. The downside of getting something wrong here is limited though IMO, so happy to rush this a bit to make the release cut tomorrow, and we can always fix things later on...
@giner I'll fix your merge conflict and set automerge now, so we get a little bit of more testing before the release cut.
There was a problem hiding this comment.
I actually need to retract my approval following some input from a watchful colleague ™️. Downloading remote binaries in a bash script is fairly sketchy from a supply chain security standpoint; at the very least I'd like us to review this with less haste.
How robust is that sha check against future code changes to this script, for example?
Perhaps the clean way is to use a different image that has grpcurl baked in?
Let me try to refactor it a bit on the next iteration (remove some verbose bash 3 compatibility stuff, split into smaller easier to read, self described pieces). If it's still hard to read after that I'll look into rewriting it Python. |
I couldn't find a good image containing all: |
|
There are two somewhat orthogonal issues for me:
I suspect in most SVs setups the info pod has direct access to things like participant admin APIs so it is highly security critical so we definitely should err on the side of being overly cautious. If we can build an image containing https://github.com/grpc-ecosystem/grpc-health-probe to get the same info, that seems like a reasonable option. I don't really know any other alternative that is well maintained and doesn't get flagged. We could write some custom python program or similar but at that point, I'd also start to somewhat question whether the complexity is worth the ROI. |
f6dea57 to
1f29dc7
Compare
84c05a4 to
f229026
Compare
|
Rebased on main and squashed |
f229026 to
c57d552
Compare
c57d552 to
7a634a5
Compare
|
Rebased on the current main |
isegall-da
left a comment
There was a problem hiding this comment.
I still think that the readability gain of doing this in python would have greatly outweighed the downsides, but I accept that you see things differently.
7a634a5 to
aa9d60d
Compare
aa9d60d to
c6cd678
Compare
Head branch was pushed to by a user without write access
c6cd678 to
17f7d93
Compare
17f7d93 to
a86ab42
Compare
Possible status values: - Scan: 0 (reachable and not lagging), 1 (lagging), 2 (unreachable). - Sequencer: 0 (reachable and not lagging), 1 (lagging), 2 (unreachable), 3 (unreachable and lagging). Also in this change: - Sort keys in the output - Refactor for readability and robustness Signed-off-by: Stanislav German-Evtushenko <ginermail@gmail.com>
Signed-off-by: Stanislav German-Evtushenko <ginermail@gmail.com>
4ea3c75 to
a81df8f
Compare
Possible status values: