hcengineering/platform · error · Error

Unable to determine the required version of Rush from ${RUSH

Error message

Unable to determine the required version of Rush from ${RUSH_JSON_FILENAME} (${rushJsonFolder}). The 'rushVersion' field is either not assigned in ${RUSH_JSON_FILENAME} or was specified using an unexpected syntax.

What it means

writeTo in foundations/server/packages/client/src/blob.ts gives up after emptyChunkRetries consecutive empty chunks while downloading a blob and throws `Empty chunk received N times for blob <name> at offset <written>/<size>`. During the retry loop it logs ctx.warn('Empty chunk received, retrying') and backs off 100ms per attempt; this Error indicates the storage backend kept returning zero bytes so the transfer aborted partway, leaving the blob at 'written' of 'size' bytes.

Source

Thrown at foundations/communication/common/scripts/install-run-rush.js:143

const RUSH_PREVIEW_VERSION = 'RUSH_PREVIEW_VERSION';
const INSTALL_RUN_RUSH_LOCKFILE_PATH_VARIABLE = 'INSTALL_RUN_RUSH_LOCKFILE_PATH';
function _getRushVersion(logger) {
    const rushPreviewVersion = process.env[RUSH_PREVIEW_VERSION];
    if (rushPreviewVersion !== undefined) {
        logger.info(`Using Rush version from environment variable ${RUSH_PREVIEW_VERSION}=${rushPreviewVersion}`);
        return rushPreviewVersion;
    }
    const rushJsonFolder = findRushJsonFolder();
    const rushJsonPath = path__WEBPACK_IMPORTED_MODULE_0__.join(rushJsonFolder, RUSH_JSON_FILENAME);
    try {
        const rushJsonContents = fs__WEBPACK_IMPORTED_MODULE_1__.readFileSync(rushJsonPath, 'utf-8');
        // Use a regular expression to parse out the rushVersion value because rush.json supports comments,
        // but JSON.parse does not and we don't want to pull in more dependencies than we need to in this script.
        const rushJsonMatches = rushJsonContents.match(/\"rushVersion\"\s*\:\s*\"([0-9a-zA-Z.+\-]+)\"/);
        return rushJsonMatches[1];
    }
    catch (e) {
        throw new Error(`Unable to determine the required version of Rush from ${RUSH_JSON_FILENAME} (${rushJsonFolder}). ` +
            `The 'rushVersion' field is either not assigned in ${RUSH_JSON_FILENAME} or was specified ` +
            'using an unexpected syntax.');
    }
}
function _getBin(scriptName) {
    switch (scriptName.toLowerCase()) {
        case 'install-run-rush-pnpm.js':
            return 'rush-pnpm';
        case 'install-run-rushx.js':
            return 'rushx';
        default:
            return 'rush';
    }
}
function _run() {
    const [nodePath /* Ex: /bin/node */, scriptPath /* /repo/common/scripts/install-run-rush.js */, ...packageBinArgs /* [build, --to, myproject] */] = process.argv;
    // Detect if this script was directly invoked, or if the install-run-rushx script was invokved to select the
    // appropriate binary inside the rush package to run

View on GitHub (pinned to 63e28dc964)

Solutions

Verify the blob in the storage backend — check its actual size/etag against the expected size; re-upload if truncated.
Retry the blob transfer; transient storage glitches can produce empty reads that succeed on a later attempt.
Delete the broken blob and regenerate it from source data, since the stored content is shorter than the advertised size.
Check storage backend health/configuration (credentials, endpoint, consistency) if multiple blobs are affected.

Defensive patterns

Strategy: retry

Try / catch

try {
  await blob.writeTo(writable)
} catch (err) {
  if (String(err).startsWith('Empty chunk received')) {
    // verify blob integrity, then retry the transfer from offset 0 with backoff
    await retry(() => blob.writeTo(writable), { retries: 3, backoffMs: 500 })
  } else throw err
}

Prevention

Verify blob size/etag against metadata before downloading; treat mismatches as corrupted data.
Retry empty-chunk failures with backoff before giving up, and always resume/verify after a failure.
Monitor the storage backend for truncated uploads or consistency problems.
Avoid deleting blobs while they are being read.

When it happens

Trigger: Reading a blob whose reported size is larger than the actual data available in storage — each read at the current offset returns an empty chunk until the retry budget is exhausted (default consecutive empty reads).

Common situations: Truncated or corrupted blob in the storage backend (upload failed mid-write); size metadata out of sync with stored content; misconfigured/inconsistent object storage (S3/MinIO) returning empty bodies; concurrent deletion of the blob mid-read.

Related errors

AI-assisted analysis of hcengineering/platform@63e28dc964 (2026-08-29). Data as JSON: /api/errors/b1a10ad76d878802. Report an issue: GitHub.