Distribution agent action messages in session details

We are using SQL 2K std with SP4 and push subscription. We have situation
between our publishing SQL server and subcribing SQL server.
Here is what I see when I double click an action message (listed as Timeout
Expired) in the listing of Distribution Agent history.
3:24 pm Initializing
3:29 pm The process is running and is waiting for a response from one of
the backend connections.
3:29 pm Timeout expired
The sequence of action took exactly five minutes. And the timeout happened
again at 3:30 pm and last until 3:35 pm. Then at 3:36, it started working
and delivered the transactions.
Can someone tell me what the 'backend connection' the distribution agent is
waiting for?
Also, during this problem period, both publishing and subscribing SQL
servers all of a sudden became activity dead meaning there is no IO and no
SQL internal activity (my perf monitor showed almost a blank screen except
cpu time), the publishing SQL server's cpu time is very minimal but the
subscribing SQL server's cpu is 60% peak and no IO. I checked the proces of
the CPU in Em and it was for replication but it had long wait time. It
looked all a sudden nobody can connect to the SQL servers. In fact, we
confirmed that users can't connect to the web application that use SQL
servers. I don't see any networking error shown in the event log in both
servers. I am assuming there is a 'disconnect' or something is 'hanging', I
like to know if the 'disconnect or hanging' caused the distribution agent to
time out or the distrbitution timeout problem caused the SQL server system to
become 'disconnect or hang'?
wingman
The replication subsystem issues commands and waits for responses. If it
doesn't get a response in a predetermined amount of time it gives the
waiting for a response from one of backend connections.
I can't explain why no-one could connect - perhaps the database or tlog was
auto-growing at that time.
Hilary Cotter
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
"Wingman" <Wingman@.discussions.microsoft.com> wrote in message
news:76F91F8B-925F-4152-9C2D-ACB9EA89FCD5@.microsoft.com...
> We are using SQL 2K std with SP4 and push subscription. We have situation
> between our publishing SQL server and subcribing SQL server.
> Here is what I see when I double click an action message (listed as
> Timeout
> Expired) in the listing of Distribution Agent history.
> 3:24 pm Initializing
> 3:29 pm The process is running and is waiting for a response from one of
> the backend connections.
> 3:29 pm Timeout expired
> The sequence of action took exactly five minutes. And the timeout
> happened
> again at 3:30 pm and last until 3:35 pm. Then at 3:36, it started working
> and delivered the transactions.
> Can someone tell me what the 'backend connection' the distribution agent
> is
> waiting for?
> Also, during this problem period, both publishing and subscribing SQL
> servers all of a sudden became activity dead meaning there is no IO and no
> SQL internal activity (my perf monitor showed almost a blank screen except
> cpu time), the publishing SQL server's cpu time is very minimal but the
> subscribing SQL server's cpu is 60% peak and no IO. I checked the proces
> of
> the CPU in Em and it was for replication but it had long wait time. It
> looked all a sudden nobody can connect to the SQL servers. In fact, we
> confirmed that users can't connect to the web application that use SQL
> servers. I don't see any networking error shown in the event log in both
> servers. I am assuming there is a 'disconnect' or something is 'hanging',
> I
> like to know if the 'disconnect or hanging' caused the distribution agent
> to
> time out or the distrbitution timeout problem caused the SQL server system
> to
> become 'disconnect or hang'?
> wingman
>
>
|||Is it a good idea to increase the query timeout from 300 seconds (default) to
maybe 600 seconds in Distribution Agent profile to accomendate any potential
long running process?
Wingman
"Hilary Cotter" wrote:

> The replication subsystem issues commands and waits for responses. If it
> doesn't get a response in a predetermined amount of time it gives the
> waiting for a response from one of backend connections.
> I can't explain why no-one could connect - perhaps the database or tlog was
> auto-growing at that time.
> --
> Hilary Cotter
> Looking for a SQL Server replication book?
> http://www.nwsu.com/0974973602.html
> Looking for a FAQ on Indexing Services/SQL FTS
> http://www.indexserverfaq.com
> "Wingman" <Wingman@.discussions.microsoft.com> wrote in message
> news:76F91F8B-925F-4152-9C2D-ACB9EA89FCD5@.microsoft.com...
>
>
|||QueryTimeout is more often used in merge replication than in transactional
replication. I normally set the inactivity level to something higher, but
this could mask other problems.
Hilary Cotter
Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html
Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com
"Wingman" <Wingman@.discussions.microsoft.com> wrote in message
news:649BE7BF-C83E-44DC-8B8A-BD5A498510CB@.microsoft.com...[vbcol=seagreen]
> Is it a good idea to increase the query timeout from 300 seconds (default)
> to
> maybe 600 seconds in Distribution Agent profile to accomendate any
> potential
> long running process?
> Wingman
>
> "Hilary Cotter" wrote: