[lixd]

What type of PR is this?

/kind bug

What this PR does / why we need it:

When a V2 task log request fails, the log panel can remain on a spinner with no way to recover. End loading on failure, show an explicit error and a reload button, and preserve the received log bytes and offset so reloading can resume without losing or duplicating output. Normal initial loading and periodic log updates use the same error handling.

Which issue(s) this PR fixes:

Found during browser regression against the OCI backend; no linked issue.

Special notes for reviewers:

• Three new unit regressions failed before the fix and pass afterward, including interruption of a partially delivered log and resuming from its saved offset. Full suite: 9 suites / 53 tests passed.
• Production build with ADD_WORKLOAD=true, Prettier, and whitespace checks passed. ESLint has zero errors and the two existing hook dependency warnings in the log component.
• Live browser regression used an isolated OCI server/Agent at backend commit 13ced73ddf30ee48cc550321383f76c9be124e8a. Stopping the temporary Agent produced a real log API ServiceUnavailable response and an explicit error panel. Restarting the Agent and clicking reload in the same dialog restored the real task log. Successful/failed task logs and node IP display also passed.
• Node SSH and Kubernetes kubectl terminal command I/O and remote PTY resize were verified through that OCI backend. The kubectl case used an existing running Kubernetes cluster; it does not qualify OCI cluster provisioning. Temporary processes, credentials and metadata were removed after testing.

Does this PR introduced a user-facing change?

Task log panels now show request failures and allow reloading logs instead of remaining on a loading spinner.

Additional documentation, usage docs, etc.:

None.