⚠ TeamCity no longer accepts the token your
options run on, so they are paused. Log out and log in with a fresh TeamCity token — they resume on their own.
GitHub login:
A GitHub personal access token is needed —
create one (classic, public_repo scope),
then paste it here. It goes into your session cookie, encrypted, and the server stores it encrypted
while an option that needs it is on.
⚠ GitHub rejected your saved token, so it was dropped and
the two options that need it were switched off — acks and the live status come from the checker's
own account meanwhile. Paste a fresh one above and switch them back on.
A JIRA Personal Access Token is needed —
create one in your ASF JIRA profile,
then paste it here. It goes into your session cookie, encrypted, and the server stores it encrypted
while auto-visa is on.
⚠ JIRA rejected your saved token, so it was dropped and
auto-visa was switched off. Paste a fresh one above and switch it back on.
Start RunAll from a pull request: switch on PR commands in ⚙ and comment /run-all.
See which tests your Apache Ignite PR actually broke —
not the ones already failing or flaky on master.
Trigger a RunAll or re-run just the blockers, right from here.
An open-source continuous-integration tool for
Apache Ignite
contributors — source on
GitHub.
It reads the project's TeamCity CI (ci2) to tell real regressions from pre-existing flaky failures.
Log in with your own TeamCity (ci2) access token. There is no shared server token: the runs you
start or cancel here go under your account.
The token is kept encrypted in your session cookie. For about an hour after your last request the
server also holds it in memory and uses it only to read from TeamCity, to pre-analyse open PRs for
everyone.
Turn on an option in ⚙ and the server stores the token encrypted until you turn the options off.
With any option other than PR commands on, it also keeps doing those reads.
Your login lasts until you log out.
No ci2 account or token yet?
Register an account on ci2, the
TeamCity of Apache Ignite.
Create an access token there: Profile → Access Tokens (“Create a token” above opens it).
The token acts with your account's rights. Reading verdicts needs only view access; Run and
Rerun need the right to run builds, “Run at top” and “Rerun at top” also to reorder the build
queue, and “Cancel my runs” to stop builds and remove them from the queue.
Pick a PR on the left, or type its number in the filter.
Didn't complete a reliable run — compilation, timeout, out-of-memory, crash or infrastructure.
A timed-out or crashed suite is shown here rather than mined for blockers (its failures would be hang cascade,
and some tests never ran). Re-run these first: the verdict below can't trust what they would have found.
These suites ran noticeably fewer tests than the same suites run on master. Tests that never
ran can't fail — a suite can look green while silently skipping coverage (a broken suite class, an excluded
package, a run that ended early). Intentional if this PR removes tests; otherwise re-run and check.
These suites hit a timeout, out-of-memory error or JVM crash, but only after running over 90% of
master's tests. Every test they failed is pre-existing or flaky (see Filtered out), so the results of the
tests they ran stand.
Posting a visa needs your JIRA Personal Access Token —
create one in your ASF JIRA profile,
then paste it here. It stays in your session cookie only, encrypted.
Each of these failed the only time it ran on this branch. One failure is not proof: re-run
their suites. A test that fails again stays a blocker, one that passes drops out.
Nothing in this run looks caused by your PR — good to go.
Failing on the current revision, with too few runs of it to tell a real break from a
flake — the passes above them ran on other code. A re-run settles it: keep failing and they become
blockers, pass and they drop out.
These tests failed, but TeamCity errors kept them from being compared with master and the
branch. They are not counted as blockers, and the checker tries them again on its own.