Philip O'Toole
|
3f5a5f076a
|
Check nodes before deprovision
|
10 months ago |
|
Philip O'Toole
|
090ff34df4
|
Revert helpers
|
10 months ago |
|
Philip O'Toole
|
4ca3420813
|
Default to legacy output for Nodes
New output can be requested via "ver=2".
|
10 months ago |
|
Philip O'Toole
|
2405063dff
|
System testing uses new form of /nodes
|
10 months ago |
|
Philip O'Toole
|
059440b7b4
|
Enable legacy during testing
|
10 months ago |
|
Philip O'Toole
|
87ccc1f21a
|
End-to-end testing of auth join
|
10 months ago |
|
Philip O'Toole
|
173120194e
|
Actually enable auth for Notify
|
10 months ago |
|
Philip O'Toole
|
097382001d
|
Join and Notify set credentials
|
10 months ago |
|
Philip O'Toole
|
41fb5d5a1a
|
Remove blank line
|
10 months ago |
|
Philip O'Toole
|
2341a77a0c
|
Allocate addresses which have never been allocated
Previously there was a very small chance that random_addr() would return
an address that had just been handed out previously. This would cause
random test failures.
See https://app.circleci.com/pipelines/github/rqlite/rqlite/3463/workflows/24f6f66c-5ac3-463d-8233-6c01b6d23a25/jobs/23213
|
10 months ago |
|
Philip O'Toole
|
e2c2a3f0d2
|
Don't wait for leader
Killing the leader, when the only other node is a read-only node, will not result in a new leader.
|
10 months ago |
|
Philip O'Toole
|
ea80fa29cf
|
Use APIProtoAddr
|
10 months ago |
|
Philip O'Toole
|
129827d092
|
Reenable a test
|
10 months ago |
|
Philip O'Toole
|
c47f364369
|
More e2e join address fixes
|
10 months ago |
|
Philip O'Toole
|
246556be9c
|
More use of TLS client
|
10 months ago |
|
Philip O'Toole
|
c52d94e26e
|
Allow test nodes to enable TLS client
|
10 months ago |
|
Philip O'Toole
|
e6fb72c5b3
|
Differentiate between "not leader" and "no leader"
Needed to handle redirected join properly
|
10 months ago |
|
Philip O'Toole
|
12c20a99c5
|
Move system testing to Raft Join and Notify
|
10 months ago |
|
Philip O'Toole
|
2da8a76060
|
Comment out auth join test for now
|
10 months ago |
|
Philip O'Toole
|
0692e138ad
|
Update E2E for Raft port join
|
10 months ago |
|
Philip O'Toole
|
ec789840c0
|
Start moving e2e tests to Raft joins
|
10 months ago |
|
Philip O'Toole
|
0d959f174e
|
Fix system-level testing of Bootstrapping
|
10 months ago |
|
Philip O'Toole
|
4e5fd30be6
|
Even less-intensive tests
|
1 year ago |
|
Philip O'Toole
|
2fff800273
|
Substantially shrink Test_MultiNodeClusterLargeQueuedWrites
|
1 year ago |
|
Philip O'Toole
|
c80ba4907f
|
Revert
|
1 year ago |
|
Philip O'Toole
|
10b003a5e0
|
Bump election timeout
Still seeing unexplained elections on CI. Let's try one more increase.
|
1 year ago |
|
Philip O'Toole
|
2ecb01b2a3
|
QueryNone may take a second to reach consistency
|
1 year ago |
|
Philip O'Toole
|
c9e97ac546
|
Comments
|
1 year ago |
|
Philip O'Toole
|
bf3baee9be
|
Tigthen up test code
|
1 year ago |
|
Philip O'Toole
|
61697967c7
|
Fix up test
|
1 year ago |
|
Philip O'Toole
|
d78d2f4d9d
|
Short circuit EOF from decompressor
The gzip NewReader() seems to hang sometimes when reopened with the old
reader. Therefore just record that we have closed, and never reopen it.
|
1 year ago |
|
Philip O'Toole
|
23fc66c50b
|
Test improvements
|
1 year ago |
|
Philip O'Toole
|
cb384a9df2
|
More test tweaks
|
1 year ago |
|
Philip O'Toole
|
d5bdaa9214
|
Tighten up helper function
|
1 year ago |
|
Philip O'Toole
|
73f9df5283
|
More end-to-end leader checks
|
1 year ago |
|
Philip O'Toole
|
f3c24cdac2
|
Give time for Store to open
https://app.circleci.com/pipelines/github/rqlite/rqlite/3300/workflows/b903b7eb-fbd7-417c-84af-fdf2189acde3/jobs/20719
|
1 year ago |
|
Philip O'Toole
|
20e6a39c1b
|
Add debug for failing test
https://app.circleci.com/pipelines/github/rqlite/rqlite/3299/workflows/34577edb-32c8-4a00-8e59-723c0e6f3272/jobs/20692
|
1 year ago |
|
Philip O'Toole
|
3315d6f91f
|
Expand end-to-end snapshot testing
|
1 year ago |
|
Philip O'Toole
|
7a8d88756c
|
WIP snapshot upgrade
|
1 year ago |
|
Philip O'Toole
|
d6aff672e2
|
Re-enable snapshot-less upgrade-from-7 testing
|
1 year ago |
|
Philip O'Toole
|
fbd6c3d987
|
Merge remote-tracking branch 'origin' into remove-ott-compression-controls
|
1 year ago |
|
Philip O'Toole
|
b9d323dc53
|
Fix test while I'm at it
https://app.circleci.com/pipelines/github/rqlite/rqlite/3269/workflows/2705ed09-1155-49ac-b320-7867341e3484/jobs/20205
|
1 year ago |
|
Philip O'Toole
|
10d2f36474
|
Remove more flag setting
|
1 year ago |
|
Philip O'Toole
|
cb989d09a2
|
Some changes from PR v2
|
1 year ago |
|
Philip O'Toole
|
76867a682f
|
Time FSM snapshot creation
|
1 year ago |
|
Philip O'Toole
|
c6dcad60f8
|
Remove more on-disk references
|
1 year ago |
|
Philip O'Toole
|
911696ce4f
|
Remove more references to on-disk testing
|
1 year ago |
|
Philip O'Toole
|
ac34308d56
|
Remove explicit on-disk testing from end-to-end
|
1 year ago |
|
Philip O'Toole
|
b931e387be
|
First pass at removing in-memory support
|
1 year ago |
|
Philip O'Toole
|
35976fc33a
|
Increase election timeout during testing
CI systems might be a bit slower.
|
1 year ago |