Closed sliekens closed 8 months ago
Can you reproduce the error after adding --driver-logging-level=65
to the command line? It may provide more insight on what happened at the driver level.
@shueybubbles no problem (I removed some portion of IPs because I don't know who else is reading this)
DRIVER:unable to get instances from Sql Server Browser on host redacted: read udp 172.23.x.x:46208->10.208.x.x:1434: i/o timeout
DRIVER:Skipping protocol tcp. Error:no instance matching 'inst01' returned from host 'redacted'
I checked with our network engineer, there is no visible traffic on port 1434. However we do see a connection on port 60000.
Do you know what it means?
1434
is the SQL Browser service port. The driver tries to query the browser service on that port to get the port number for inst01
.
If you know the port number for inst01
you could change your connection string to use it directly instead of using the instance name, it will speed up the connection a bit.
Alternately try increasing the dial timeout.
Thanks. I was told the port for inst01
is 50001. Do you have any insight what the connection to port 60000 could be?
How would I increase the dial timeout? Is it the -l
option?
I don't know what 60000 would be.
-l
is the dial/connect timeout
Just gonna close this as I'm waiting for krb5 support to switch from old sqlcmd to sqlcmd-go. It was probably an environmental issue anyway.
Hello,
I intermittently get an error with sqlcmd v.1.2.1 when connecting to an instance in a high availability cluster.
This sounds to me like an issue with MultiSubnetFailover. I removed
-M
from my sqlcmd command because it has been deprecated, but I didn't have this problem with the ODBC version of sqlcmd (with-M
).My full command is something like:
OS: Debian 11 running inside Docker
Thanks for your attention.