unstable SQL service

unstable SQL service

Post by Darryl Fryat » Sat, 09 Aug 1997 04:00:00



Hi,

Please help with the following problem:

8:30am          SQLServer fine, no problems
9:15am          Users Logging on....
9:20am          Users Complain, they cannot connect to SQL....
9:25am          I cannot connect using enterprise manager or any other tool......
9:30am          Try to shut down service using SQL Service Manager....SQL Executive
shuts down,  SQLService says "Service is           stopping......"
9:45am          SQLService still says "Service is stopping......"
9:50am          Try using NET STOP SQLSERVICE........message says something like
"Cannot control service in its present             state"..........very worrying!!!
9:55am          Shut Down NT box, power off, reboot..........everything
fine................until next time!!!!!!

Any ideas?

Darryl Fryatt   Database Manager
                Sub-Contract UK Ltd
                Basingstoke

Tel No: 44+(01256) 840101

 
 
 

unstable SQL service

Post by Neil Pik » Sun, 10 Aug 1997 04:00:00


Darryl,

 Check the SQL errorlog for messages - it's probably having internal AV's or
something like that.

 Neil Pike MVP/MCSE
 Protech Computing Ltd (MS Solution Provider)
 Using Virtual Access 4.01 build 232 (32-bit) on NT 4.0 SP3

 
 
 

1. Unstable sql server - locking/blocking/1222 errors help pls

Hi, we have a single sql 2000 server with 4 CPUs. We are finding that each
monday the server critically slows down. Some mondays it is due to blocking,
other mondays we find there are a fair few deadlocks occuring. The server is
fine for the rest of the week. In relation to deadlocks, most of the
deadlocks are specific to two tables. I have set flag 1204 and have logged
the deadlock info to the sql log. I am finding that the locks are occuring
on the index i.e. I only get 'modes' of type range-X-X, range-S-U etc for
instance. So these represent range locks on indexes. This was accompanied by
a fair few 1222 errors (lock timeout errors). The only difference between
the friday and the monday is that we have a maintainance plan running that
drops/recreate indexes for the databases.

It seems all fingers are pointing towards the indexes. I have checked
everything else. Perf mon is running fine. I ran profiler and it mainly only
comes up with deadlock issues. The profiler log also shows the duration of
committed queries taking upto 200 seconds. Maybe I should do some sort of
index analysis. I am not clear on the steps to take. I might use sql
statistics to see which colums are most used on the tables that are involved
in deadlocks and reindex them according to columns with more unique
values/most accessed. Due to the locking, it seems that the indexes are too
clustered together and that they should be more spread out as im getting a
lot of range-x-x locks occuring.

I ran the index tuning wizard through a sample 5 min profiler trace I
generated, but it made no alterations to the indexing it was alreading
using.

Would anyone have any suggestions? Id be much appreciated!

Cheers,
James

2. little VBA problem but ... too big for me

3. SQL Server 7 unstable on Windows 98

4. D2.01/Sybase: App hangs on duplicate key violation

5. Sql Server 2000 seems unstable

6. Transferring Sybase System 11 database to MS SQL

7. SQL server 7.0 unstable.

8. Micro-ISIS for Windows

9. NT/SQL7.0 unstable..?

10. SQL2000->Access2000 Unstable Performance

11. unstable performance

12. server /database unstable