I have observed the following behavior in different 11.0.1 versions, for which I don't have any explanation:
As soon as a sequential scan on such a table starts all new connections to the database server are accepted but NOT processed. This means all other db-clients are waiting the 2-4 minutes before they can interact with the database. In this case it doesn't matter which of the 4 databases the client wants to access. So even connections to one of the other databases (not doing the sequential scan right now) will be frozen. The same for dbconsole it will also wait until the sequential scan finishes.
In the request log I just see that the connections are accepted but no single statement (even not the if ... select @@version ...) which I normally see as the first statement will be executed until the sequential scan is finished.
Have I configured anything wrong? Is this a desired behavior or a bug?
asked 01 Dec '10, 09:06
I am having a problem that is very similar to this, all connections wait. I am not sure what query they wait for but the read-IO is high and constant on the DB for the entire duration of the wait period (in my case this DB is ~300gb and the wait time is about 30 mins, IO about 2mb/s).
I cannot easily repeat the problem on demand as I dont know what query is causing it.
There is however only 1 table in this DB, a table of blobs with an ID.
Would it be useful for the future / Sybase product to try repeat this issue or to get stats on the table before refreshing them (which seems to be a possible solution, given the above and the deletes).
It is very serious because of the time involved in this case (~30 mins) but also because Sybase Central cannot itself connect and hence it is not possible to attempt to trouble shoot from within Sybase Central.