We're running a 2014 SQL Server. For the past couple of years, the server generates memory dumps with no rhyme or reason, every 2-7 days, locking up most processes, and forcing us to restart the SQL service to get it going.
Here's snippet of the error file:
Exception Address = 00007FFAD7FCBAA2 Module(combase+000000000003BAA2)
Exception Code = c0000005 EXCEPTION_ACCESS_VIOLATION
Access Violation occurred reading address 00000000D8063158
Input Buffer 74 bytes -
Anything else we should explore?
Or is our ONLY option to have Microsoft analyze the .MDMP
file(s) ?
Points to mention:
We have it patched to the latest Service Package, with the latest Cumulative & GDR updates. Hasn't made a difference
Microsoft SQL Server 2014 (SP3-CU4-GDR) (KB5014164) - 12.0.6439.10 (X64)
There's no specific SQL statements/stored procedures that correlate to when these memory dumps occur.
The SQL database files run on a seperate Hard drive than the OS. It's a 1TB drive with 250GB free at any given time.
The Machine is a Windows Server 2012 R2, Virtual Machine
CHKDSK
does not find any errors on the OS & Data drives.DBCC checks go through without issues.
OPENQUERY
&OPENROWSET
functions are used quite frequently, using the latestACE
&ODBC
drivers.If I query sys.dm_server_memory_dumps, here's the record count by date:
creation_time QTY 2022-09-13 1 2022-09-09 38 2022-09-07 11 2022-09-06 21 2022-09-04 1 2022-08-25 2 2022-08-24 11 2022-08-23 7 2022-08-22 1 2022-08-21 1 2022-08-17 9 2022-08-09 7 2022-07-29 1 2022-07-27 1 2022-07-22 1