Project History Sharded Blobs Collection

projectHistoryShardedBlobs holds the blob records that did not fit into projectHistoryBlobs: when the bucket for a hash prefix already has 8 records there, the next ones go here. A project has at most 16 documents in it, one per first hex digit of the hash. There is no Mongoose schema; history-v1 writes it in services/history-v1/storage/lib/blob_store/mongo.js (layout described at mongo.js:1-16).

Fields

Field	Type	Meaning	Source
_id	Binary	13 bytes: the history id of the project, a 0, and the shard digit, read as hex (<history id>0<digit>). The shard digit is the first hex digit of the hash.	mongo.js:439-453
blobs	Object	Blob records grouped in buckets. The key of a bucket is hex digits 2 to 4 of the hash.	mongo.js:439-443
blobs.<bcd>	Array of Object	The records of the bucket. The code puts no limit on its size.	mongo.js:399-407
blobs.<bcd>[].h	Binary	The hash, 20 bytes, as in projectHistoryBlobs.	mongo.js:472
blobs.<bcd>[].b	Number	Byte length of the content.	mongo.js:473
blobs.<bcd>[].s	Number	Length of the content as text; null when the content is not editable text.	mongo.js:474

Who Writes It

Code	What it does
services/history-v1/storage/lib/blob_store/mongo.js:399-407	insertRecordSharded: upsert with $addToSet into the bucket. Called by insertBlob when the bucket in projectHistoryBlobs is full (mongo.js:387-389).
mongo.js:75-103	clone: copies the project’s shards to another history id, for history-v1 POST /projects/:project_id/clone, which nothing in web calls.
mongo.js:413-422	deleteBlobs: deletes all shards of the project together with its projectHistoryBlobs document, on history-v1 DELETE /projects/:project_id.

Who Reads It

Code	What it does
mongo.js:144-157	findBlobSharded: one blob, looked up only when its bucket in projectHistoryBlobs has 8 records and does not hold it (mongo.js:127-131).
mongo.js:217-260	findBlobsSharded: the blobs findBlobs did not find in projectHistoryBlobs (mongo.js:202-206).
mongo.js:285-306	getProjectBlobs: all shards of a project.
mongo.js:337-361	getProjectBlobsBatch: all shards of several projects, for the blob stats route.

Indexes

Only the default _id index. No migration in tools/migrations/ adds one.

Notes

• The shards of one project sit next to each other in _id order, so the code finds them with an _id range from <history id>00 to <history id>0f (mongo.js:75-80, 287-295).
• The collection only appears once some project has more than 8 blobs with the same 3-digit hash prefix, so a small instance may not have it at all.