BAD_PAGE_PROCESS Wait Type in SQL Server

---

BAD_PAGE_PROCESS is recorded by the background suspect page logger, the process that records 823 and 824 I/O errors into msdb.dbo.suspect_pages, as it throttles itself to run no more than once every five seconds. You will see at most one wait per 823/824 error.

That makes this the rare wait type whose count is the whole message: each occurrence corresponds to a corruption event.

Is It a Problem?

Any meaningful presence is a five-alarm signal, not a performance discussion. Frequent BAD_PAGE_PROCESS waits mean repeated 823/824 errors, which means either widespread corruption or a workload repeatedly hitting the same corrupt pages. Either way, you are in incident territory: data is unreadable somewhere, and the underlying cause is usually the I/O subsystem.

This is one wait where “filter it out” is exactly wrong.

Common Causes

• Storage subsystem faults corrupting pages (controller, firmware, cabling, SAN issues).
• A corrupt hot page being re-read constantly by the workload.
• Memory faults writing bad pages that later fail checksum on read.

What To Do

1. Check msdb.dbo.suspect_pages immediately (our Get-SuspectPages script and the health check both surface this) and the error log for the 823/824 details.
2. Run DBCC CHECKDB on the affected databases to scope the damage.
3. Repair from backups where possible: page-level restore for isolated pages, full restore for widespread damage. Do not reach for REPAIR_ALLOW_DATA_LOSS before exhausting restore options.
4. Involve the storage team in parallel; corruption that keeps recurring will not stop until the hardware cause does.

How To See It

Rank waits with Get-WaitStatistics, but do not wait for it to rank: suspect_pages and the error log are the primary alarms, and our health check reviews both.

---