New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 676436 link

Starred by 1 user

Issue metadata

Status: Assigned
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: All
Pri: 3
Type: Bug



Sign in to add a comment

Download history queries are reporting bad and duplicate IDs

Project Member Reported by asanka@chromium.org, Dec 21 2016

Issue description

DownloadDatabase logs metrics on when history records are dropped. Based on the metrics we see there are a non-negligible number of events where:

 * The download record on disk had a negative ID. Only positive IDs should ever be getting written to the DB.

 * There are multiple records with the same ID. There should be no cases where a duplicate ID is written to the DB.

The UMA in question is:

  Download.DatabaseRecordDropped

The UMA is recorded in:

  void DownloadDatabase::QueryDownloads(std::vector<DownloadRow>* results)

We should have a closer look and figure out what's going on. If the ID field is unreliable, then the other fields are even less reliable.
 
Cc: nyquist@chromium.org
Owner: xingliu@chromium.org
Status: Assigned (was: Untriaged)
Notes:
- ID set based on DB query and incrementing largest ID.
- Happening according to UMA, figure out if UMA issue or actual DB bug.
- Investigate source of UMA data.
Interesting. I am guessing that multiple files are often downloaded at the same time, so we should ensure that the query-increment-store pattern works correctly (check for races, etc.)

Comment 3 Deleted

Comment 4 Deleted

Sign in to add a comment