November 17, 2016 at 2:12 pm
We have a 2008 DB, with 3 client servers all on the same subnet as the sql server.
2 machines connect just fine, 1 machine throws this error on the initial login attempt:
Unable to complete login process due to delay in opening server connection
Subsequent logins are work fine.
We've gone through some of the other blog posts, but nothing seems to work.
pings come back in less than 1ms
traceroutes only have 1 hop
the connection is over a non standard port, which is open and the sql browser port is open as well (tested with portqueryui)
We have the issue if use the SQL server FQDN or the IP address.
Honestly, I'm not sure where to go next.
November 17, 2016 at 2:53 pm
an0nemus (11/17/2016)
We have a 2008 DB, with 3 client servers all on the same subnet as the sql server.2 machines connect just fine, 1 machine throws this error on the initial login attempt:
Unable to complete login process due to delay in opening server connection
Subsequent logins are work fine.
We've gone through some of the other blog posts, but nothing seems to work.
pings come back in less than 1ms
traceroutes only have 1 hop
the connection is over a non standard port, which is open and the sql browser port is open as well (tested with portqueryui)
We have the issue if use the SQL server FQDN or the IP address.
Honestly, I'm not sure where to go next.
Quick questions, any conflicting entries in the host file on the troubled machine? Have you flushed the dns cache? Have you tried connecting with sqlcmd?
😎
November 18, 2016 at 6:38 am
1) host file has just the 'usual' entries
2) problem server has been rebooted
3) I'm not sure what sqlcmd is, but SQL Server Mangement studio will also have issues, though less frequently
November 18, 2016 at 8:11 am
an0nemus (11/18/2016)
1) host file has just the 'usual' entries2) problem server has been rebooted
3) I'm not sure what sqlcmd is, but SQL Server Mangement studio will also have issues, though less frequently
sqlcmd is a command line utility for SQL Server, the thing is that there is no cached host entries or credentials like in SSMS so if a connection works from there, then the problem could possibly be somewhere in an application connection cache. Further, have you checked the NIC on the troubled server, wouldn't be the first time that would be the cause. Wireshark on a linux box would be the default tool for me to analyse that kind of a problem.
😎
November 18, 2016 at 1:26 pm
I've seen this happen recently on a Web/WPF app where the SQL Instance was using SQL Express;
as a resource conservation attempt, the express version defaults databases to auto close,and some applications will timeout while the database is being "opened" again.
maybe that might be the culprit? especially if the problem applciaiton is using a "default" database where the database was closed.
SELECT *
FROM sys.databases
WHERE [databases].[is_auto_close_on] = 1
Lowell
November 23, 2016 at 12:53 pm
Just curious if there's any chance that the clock on the problem server is not within a minute or two of the SQL Server clock. If there's any portion of the connection that might measure timeout by time on one server vs time on another, then this might be able to occur...
Steve (aka sgmunson) 🙂 🙂 🙂
Rent Servers for Income (picks and shovels strategy)
November 23, 2016 at 1:08 pm
sgmunson (11/23/2016)
Just curious if there's any chance that the clock on the problem server is not within a minute or two of the SQL Server clock. If there's any portion of the connection that might measure timeout by time on one server vs time on another, then this might be able to occur...
Interesting thought but I wouldn't have thought that this could be the problem unless there are some strange "propitiatory" things in place.
😎
November 23, 2016 at 1:48 pm
Eirikur Eiriksson (11/23/2016)
sgmunson (11/23/2016)
Just curious if there's any chance that the clock on the problem server is not within a minute or two of the SQL Server clock. If there's any portion of the connection that might measure timeout by time on one server vs time on another, then this might be able to occur...Interesting thought but I wouldn't have thought that this could be the problem unless there are some strange "propitiatory" things in place.
😎
Doesn't really require much in the way of propitiation... just one ODBC driver developer too lazy to call the system time because it happens to be right there in his "retrieved data", so... gosh darn it, he's gonna use it, on account of the need to not waste ANY time getting a connection or shutting one down, and besides, who on god's green earth doesn't set their time correctly???
Yes, that's all sarcasm, and yes, seen it happen...
Steve (aka sgmunson) 🙂 🙂 🙂
Rent Servers for Income (picks and shovels strategy)
Viewing 8 posts - 1 through 8 (of 8 total)
You must be logged in to reply to this topic. Login to reply