[
https://issues.apache.org/jira/browse/HDDS-16548?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18127790#comment-18127790
]
Siyao Meng commented on HDDS-16548:
-----------------------------------
Verification run complete. The Specula pipeline ran all phases against the
pinned commit: code analysis, TLA+ specification generation, harness and trace
collection, trace validation plus TLC model checking, bug confirmation with the
adversarial Challenger debate, and severity classification.
h3. Run environment
{noformat}
Ozone commit: 916aa1d9949cbb10f2750a40c8704340893131a2
Specula: v1.2.0-32-g54fe3a93 (with local changes)
Agent/model: claude-code, Claude Opus 5.5 (1M context)
Effort: medium
TLC limits: 28G memory, 8 workers
Total: 3h 25m agent time, 181.8M tokens, about $102 (estimate)
{noformat}
h3. Result
12 candidate findings came out of model checking (10) and code review (2), and
all 12 were investigated individually.
||Disposition||Count||
|Reproduced|9|
|Masked (defect present, harm prevented by another component today)|3|
Some findings share one fix and are filed as one issue: two dead node handling
findings (HDDS-17067), two Recon rebuild findings (HDDS-17061), and the health
task retry loop, which is filed with the pipeline sync loop (HDDS-17071). Each
issue was then validated a second time by hand against the pinned commit: a
regression test was added to an existing suite and shown to fail without the
fix and pass with it, the bug was confirmed still present on apache master, and
the fix patch, reproduction and report were reviewed independently before
filing. The three masked findings are filed as well, because each leaves
Recon's (and for two of them SCM's) view wrong until the next periodic full
container report. One further defect of the same kind as HDDS-17071 was found
during validation and is filed as HDDS-17065.
h3. Filed
||Issue||Priority||Summary||
|HDDS-17064|Major|Full container report carried by a failed heartbeat is not
resent until the next periodic report, hiding volume failure losses for up to
two hours|
|HDDS-17067|Major|Registration does not refresh a known datanode's heartbeat
time, so the dead node handler keeps or erases its replicas wrongly|
|HDDS-17066|Major|Recon removes the replicas of a datanode stopped for
maintenance when it has not yet learned IN_MAINTENANCE from SCM|
|HDDS-17071|Major|Recon PipelineSyncTask stops after one failed cycle and
ContainerHealthTask retries a failed sweep without waiting|
|HDDS-17062|Minor|ICRs put back after a failed datanode heartbeat are sent
after the newer full container report that replaced them|
|HDDS-17063|Minor|ICR sent after the volume failure full container report
re-adds the lost replica in Recon and SCM|
|HDDS-17060|Minor|Recon keeps a container CLOSED after SCM moves it from
DELETING or DELETED back to QUASI_CLOSED|
|HDDS-17061|Minor|Recon drops the replicas of a DELETED container when it moves
it back to SCM's live state|
|HDDS-17070|Minor|Recon does not follow container suppression changes made in
SCM after it has added the container|
|HDDS-17069|Minor|Recon's per datanode open container count drifts from the
containers that are actually open|
|HDDS-17065|Minor|Recon ContainerSizeCountTask stops for good after one failed
run, freezing the container size distribution until restart|
Priorities follow the manual validation, not the pipeline's severity classes.
Each issue has a proposed fix patch and a standalone reproduction test attached.
h3. Not filed
* Part of a code review finding: a replica with an out of range EC index
aborting the whole health sweep. Both report paths already reject such a
replica, so it cannot reach the sweep.
* Part of a code review finding: drift in Recon's replica history and heartbeat
maps. Not reproduced; only the open container count part is filed (HDDS-17069).
h3. Reproduce
{code:none}
specula run --agent=claude-code --effort=medium --model='claude-opus-5-5[1m]' \
--max-parallel=2 --enable-reviews --confirm-debate \
--tlc-memory-limit=28G --tlc-worker-limit=8 \
"recon-container-reports|apache/ozone|Java|Use the target-specific guidance"
{code}
h3. Caveats
* The SCM side of HDDS-17064, HDDS-17067, HDDS-17062 and HDDS-17063 is argued
from the shared datanode and SCM code and was not run against SCM. The
reproductions exercise the Recon endpoint.
* Several reproductions control timing or inject the fault that starts the
scenario: a heartbeat RPC failure (HDDS-17062, HDDS-17064), a failed volume
(HDDS-17063, HDDS-17064), a latch in the dead node handler (HDDS-17067), an SCM
RPC or Recon DB failure (HDDS-17071, HDDS-17065), and a failed container add
(HDDS-17061). HDDS-17066 was also run on a mini cluster with only the
maintenance admin operation and a datanode shutdown.
* HDDS-17067 changes SCM behaviour: a known datanode that registers again
leaves STALE or DEAD at the next health check instead of after its first
heartbeat.
* HDDS-17062, HDDS-17063 and HDDS-17064 change {{StateContext}}, as does the
open pull request for HDDS-16723 (apache/ozone#11423); the main code of all
three applies cleanly on its head commit.
* The patches were built and tested on the pinned commit and each applies to
master at d77edaa6c15. Some pairs touch the same test file and need a trivial
rebase for whichever lands second: HDDS-17060, HDDS-17061 and HDDS-17070
({{TestReconSCMContainerSyncIntegration}}), HDDS-17066 and HDDS-17067
({{TestReconNodeManager}}), HDDS-17062 and HDDS-17064
({{TestHeartbeatEndpointTask}}).
* The patches are proposals for review.
Generated with Specula (Claude Opus 5.5).
> [Specula] Verify recon-container-reports (Apache Ozone master @916aa1d)
> -----------------------------------------------------------------------
>
> Key: HDDS-16548
> URL: https://issues.apache.org/jira/browse/HDDS-16548
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Siyao Meng
> Assignee: Siyao Meng
> Priority: Minor
> Labels: help-wanted, specula
> Attachments: recon-container-reports.guidance.md
>
>
> This is a runnable Apache Ozone formal-verification target for the Specula
> pipeline (TLA+ model checking plus trace validation against the real code).
> It has not been run yet. Anyone can pick it up: run Specula against the
> pinned commit, then report findings on this issue. Part of umbrella
> HDDS-15926.
> h3. Target
> {{recon-container-reports}}. Scope and code entry points are in the attached
> guidance file [^recon-container-reports.guidance.md].
> h3. Run environment
> {noformat}
> Apache Ozone commit: 916aa1d9949cbb10f2750a40c8704340893131a2
> Specula: https://github.com/specula-org/Specula (v1.1.0 or later)
> Toolchain: JDK 21, ~28 GB RAM for TLC, an LLM coding agent
> (claude-code or codex adapter)
> Runtime: hours per target; needs an agent API budget
> {noformat}
> h3. How to run
> # Clone Specula and install per its README; clone apache/ozone and check out
> the pinned commit above.
> # Place the attached guidance file as the target's .prompt-extra.md (see the
> Specula README for target layout).
> # Run:
> {code:none}
> specula run --agent=claude-code --effort=medium --keep-original
> --max-parallel=2 \
> --enable-reviews --confirm-debate --tlc-memory-limit=28G
> --tlc-worker-limit=8 \
> "recon-container-reports|apache/ozone|Java|Use the target-specific
> .prompt-extra.md"
> {code}
> # Deliverables land under the run's {{.specula-output/}}:
> {{confirmed-bugs.md}}, {{bug-severity.md}}, {{summary.md}}, and
> {{confirmation/<id>/investigation.md}}.
> h3. Reporting results
> Comment here with the outcome (0 findings is still useful: it records the
> target as covered at this commit). For each confirmed bug, file a separate
> Ozone Bug and link it to this issue with the "Testing" link type ("Testing
> discovered").
> h3. Notes
> * The attached guidance file names an earlier commit ({{20e6a1c}}); use the
> commit above. Code entry-point line numbers in the guidance were captured at
> an earlier commit and may have shifted on this head; treat them as
> approximate. Specula re-analyzes the source, so the run does not depend on
> them.
> * Recon-building targets need a local Specula patch to skip Maven {{target/}}
> out-of-tree symlinks in {{snapshotlib._validate_source_tree}}, else the run
> crashes at finalize.
> * effort=medium is the tested default for large targets; effort=high can hit
> per-turn timeouts on the biggest ones.
> * Findings are proposals, to be confirmed by human review and testing before
> any fix is merged.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
