Hi everyone,
We’ve just released NServiceBus.SqlServer 6.0.1, 5.0.3, and 4.3.2.
Fixed bug
#624 Contention on SQL Server latches for clustered index prevents transport from scaling
How to know if you are affected
Users reaching the upper limit of SQL server resources that want to scale the server up.
Symptoms
Scaling up SQL server that runs at high CPU levels (~75%) results in minimal (if any) increase in overall system throughput. Repeated execution of wait time statistics query:
SELECT session_id, wait_type, wait_time, wait_resource FROM sys.dm_exec_requests
WHERE session_id > 50 and wait_type = 'pagelatch_ex'
shows a single page (5:1:4144 in the listing below is just an example) contributing the most to the wait times on pagelatch_ex:
session_id wait_type wait_time wait_resource
---------- ---------------------- --------------- ------------------------
60 PAGELATCH_EX 20 5:1:4144
75 PAGELATCH_EX 12 5:1:4144
79 PAGELATCH_EX 40 5:1:4144
80 PAGELATCH_EX 25 5:1:4144
81 PAGELATCH_EX 31 5:1:4144
82 PAGELATCH_EX 35 5:1:4144
84 PAGELATCH_EX 8 5:1:1234
87 PAGELATCH_EX 11 5:1:7867
89 PAGELATCH_EX 5 5:1:3321
71 PAGELATCH_EX 3 5:1:2188
Should you upgrade immediately
Users affected should upgrade immediately. Changes introduced result in less resource consumption. As a result, non-affected users are advised to upgrade during the next maintenance window to take advantage of better performance.
Users can perform the upgrade one endpoint at a time, starting with the high throughput parts of the system first. Index upgrade guide describes details of the procedure.
Where to get it
You can install the new versions from NuGet.
With thanks,
The team in Particular
Please read our release policy for more details. Follow @ParticularNews to be notified of new releases and bug fixes.