-
Notifications
You must be signed in to change notification settings - Fork 96
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
test(customd2): add custom_d2/workload1 test configurations #8951
test(customd2): add custom_d2/workload1 test configurations #8951
Conversation
test-cases/longevity/longevity-aws-custom-d2-workload1-multidc.yaml
Outdated
Show resolved
Hide resolved
jenkins-pipelines/oss/longevity/longevity-aws-custom-d2-workload1-multidc-big.jenkinsfile
Outdated
Show resolved
Hide resolved
bdf009f
to
831b632
Compare
- >- | ||
latte run --tag latte --duration 180m --request-timeout 60 --retry-interval '2s,10s' | ||
--sampling 5s --threads 30 --connections 3 --concurrency 180 --rate 15000 -P offset=0 | ||
--function custom -P row_count=50100100 -P codes="\"T13F1\"" -P print_applied_func_names=2 --consistency=QUORUM |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
why in some command we specify --consistency=QUORUM
and in some we don't ?
I would guess that the default is LOCAL_QUORUM
I guess there is some assumption here that one remember by heart codes for queries...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Have you read comment on line 139?
It directly influences your question.
First command out of 5 writes to it's own region, other 4 loaders of the first region write to per-dc tables of other 4 regions. Such actiosn requires usage of non-local QUORUM for these other 4 loaders
.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In places where we do not specify CL
we depend on the default value which is local_quorum
.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've read the comment, the part about why CL usage need to be different, is not easily deduced from that comment. one need a clear mental picture of the whole setup in mind, and that really hard in this case.
maybe we should consider adding some overall description on qa-internal repo, and point to it.
I'm guessing a diagram for this case might be helpful.
831b632
to
305fc1f
Compare
95090f1
to
f130457
Compare
f130457
to
fe4c75b
Compare
60 nodes scenario passed using fix from Asias (scylladb/scylladb#21207): |
ae01500
to
d36bcc5
Compare
The latest state of the 5dcs was tested here: Using the unmerged PR (via BYO) needed for the enterprise versions: https://github.com/scylladb/scylla-enterprise/pull/5039 |
The current configuration is suitable for current Scylla master versions only. Following Scylla PR backports are needed for any other Scylla version than current master: |
SCT_GCE_IMAGE_DB=image SCT_GCE_DATACENTER="us-east1 us-west1 eu-north1" SCT_SCYLLA_REPO='http://downloads.scylladb.com.s3.amazonaws.com/rpm/centos/scylla-2021.1.repo' ./sct.py lint-yamls -b gce -i '3dcs,24h-multidc,large-cluster,counters-multidc,cdc-8h-multi-dc' -e 'docker,azure,shutdown' | ||
OUT=$(($OUT + $?)) | ||
|
||
echo "multi dc yamls with 5 regions" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
oh my...
I really hate this shell script
one day we should move it into python code, to speed things up, and make it a bit less annoying
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Agree
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
you have one unittest failure, cause of the new scylla configuration:
|
d36bcc5
to
a563c31
Compare
First configuration is "small" - 3 DCs, 3 nodes per each. Second configuration is "big" - 5DCs, 12 nodes per each. Ref: #19131
a563c31
to
e3bdd1a
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
First configuration is
small
- 3 DCs, 3 nodes per each.Second configuration is
big
- 5DCs, 12 nodes per each.This
big
setup never reached theTest start
stagebecause of the very slow bootstrap process in a multi-dc environment.
Ref: #19131
Testing
PR pre-checks (self review)
backport
labelsReminders
sdcm/sct_config.py
)unit-test/
folder)