Deadlocking on indexes?

It seems that a common problem we have is deadlocks occurring when one
stored proc is updating a row and another store proc is running a read query
against the same table. They'll often get deadlocked on two of the indexes
for the table.
For example:
SP1 is updating a row in TableA, causing it to get an exclusive lock on
TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
attempts to get a shared lock on TableAIndex1.
Deadlock!
Of course, I can fix this by lowering the isolation level in SP2 to "read
uncommitted", but I'd rather not.
I can normally fix deadlocks by changing the order in which locks are
acquired, but I don't know how to influence the order in which locks on
indexes are acquired. Is it controlled by the order in which the fields are
listed in a select or an update statement? If I standardize the order in
which individual fields are listed whenever I perform a select or update
from a given table, will that effect thr order in which index locks are
acquired?
Thoughts?
Joel
The order in which the columns appear in the query are not tied directly to
how the engine decides to take locks. Why is the query trying to use two
indexes on the same table? Is this two separate queries or is it doing
index intersection? If it is the latter you might want to see if adding the
column to the first index will solve the problem.
Andrew J. Kelly SQL MVP
Solid Quality Mentors
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel
|||One way to solve this would be to put a copy of the same Select statement
used in the SP1 for the update at the beginning of the SP2. This way, SP2
will have to first acquire the shared lock in the same order as SP1; hence
solving (I hope!) your deadlocking problem.
It has been a long time since the last time that I had to work on a locking
problem but I remember a suggestion whose idea was to put at the beginning
of a SP one or more queries with the purpose of acquiring a list of locks in
the same order as for the other SPs. At first, you might think that making
these queries will have some negative impact on the overall performance but
don't forget that most (if not all) of this stuff must be read from the I/O
and put into memory/buffer anyway. Acquiring all the locks in the exact
right order is more important than to save a few I/O or a few CPU cycles.
Don't know if there is a better way of solving this kind of problem.
Sylvain Lafontaine, ing.
MVP - Technologies Virtual-PC
E-mail: sylvain aei ca (fill the blanks, no spam please)
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel
|||Joel
In addition to others , if you run SQL Server 2005 , let DTA (Tunning
Advisor) to reccomed you what indexes are missed.
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel
|||Joe,
many things just have said by others like replications, keep only used index,
etc, but I've said one more: FILLFACTOR.
If you have a very busy OLTP database, you can controll lock at least to
minimun with appropriate fillfactor.
I administrate many OLTP servers anda database and after read many articles
and try to find a baseline to controll lock. Finally, I've used these
templates with excelent result : Recreate every clustered index with
fillfactor 80% and recreate every nonclustered index with 90%.
You can increase or decrease this value that depends if you have updated
tables that need more, but stay with a little less in clustered than
noclustered.
Try this !
Kris
Joel Lyons wrote:
>It seems that a common problem we have is deadlocks occurring when one
>stored proc is updating a row and another store proc is running a read query
>against the same table. They'll often get deadlocked on two of the indexes
>for the table.
>For example:
>SP1 is updating a row in TableA, causing it to get an exclusive lock on
>TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
>SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
>attempts to get a shared lock on TableAIndex1.
>Deadlock!
>Of course, I can fix this by lowering the isolation level in SP2 to "read
>uncommitted", but I'd rather not.
>I can normally fix deadlocks by changing the order in which locks are
>acquired, but I don't know how to influence the order in which locks on
>indexes are acquired. Is it controlled by the order in which the fields are
>listed in a select or an update statement? If I standardize the order in
>which individual fields are listed whenever I perform a select or update
>from a given table, will that effect thr order in which index locks are
>acquired?
>Thoughts?
>Joel
Message posted via droptable.com
http://www.droptable.com/Uwe/Forums.aspx/sql-server/200804/1
|||Wow! Those are some great ideas! I'll look into them further. Thank you.
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel
|||Here is, IMHO, the bible for deadlock troubleshooting:
Kevin G. Boles
Indicium Resources, Inc.
SQL Server MVP
kgboles a earthlink dt net
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel
|||slip of the finger. here is the link:
http://blogs.msdn.com/bartd/archive/2006/09/09/Deadlock-Troubleshooting_2C00_-Part-1.aspx
Kevin G. Boles
Indicium Resources, Inc.
SQL Server MVP
kgboles a earthlink dt net
"TheSQLGuru" <kgboles@.earthlink.net> wrote in message
news:8aSdnTZNUZEe02vanZ2dnUVZ_jWdnZ2d@.earthlink.co m...
> Here is, IMHO, the bible for deadlock troubleshooting:
>
> --
> Kevin G. Boles
> Indicium Resources, Inc.
> SQL Server MVP
> kgboles a earthlink dt net
>
> "Joel Lyons" <JoelL@.novarad.net> wrote in message
> news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
>
sql

Deadlocking on indexes?

It seems that a common problem we have is deadlocks occurring when one
stored proc is updating a row and another store proc is running a read query
against the same table. They'll often get deadlocked on two of the indexes
for the table.
For example:
SP1 is updating a row in TableA, causing it to get an exclusive lock on
TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
attempts to get a shared lock on TableAIndex1.
Deadlock!
Of course, I can fix this by lowering the isolation level in SP2 to "read
uncommitted", but I'd rather not.
I can normally fix deadlocks by changing the order in which locks are
acquired, but I don't know how to influence the order in which locks on
indexes are acquired. Is it controlled by the order in which the fields are
listed in a select or an update statement? If I standardize the order in
which individual fields are listed whenever I perform a select or update
from a given table, will that effect thr order in which index locks are
acquired?
Thoughts?
JoelThe order in which the columns appear in the query are not tied directly to
how the engine decides to take locks. Why is the query trying to use two
indexes on the same table? Is this two separate queries or is it doing
index intersection? If it is the latter you might want to see if adding the
column to the first index will solve the problem.
--
Andrew J. Kelly SQL MVP
Solid Quality Mentors
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel|||One way to solve this would be to put a copy of the same Select statement
used in the SP1 for the update at the beginning of the SP2. This way, SP2
will have to first acquire the shared lock in the same order as SP1; hence
solving (I hope!) your deadlocking problem.
It has been a long time since the last time that I had to work on a locking
problem but I remember a suggestion whose idea was to put at the beginning
of a SP one or more queries with the purpose of acquiring a list of locks in
the same order as for the other SPs. At first, you might think that making
these queries will have some negative impact on the overall performance but
don't forget that most (if not all) of this stuff must be read from the I/O
and put into memory/buffer anyway. Acquiring all the locks in the exact
right order is more important than to save a few I/O or a few CPU cycles.
Don't know if there is a better way of solving this kind of problem.
--
Sylvain Lafontaine, ing.
MVP - Technologies Virtual-PC
E-mail: sylvain aei ca (fill the blanks, no spam please)
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel|||Joel
In addition to others , if you run SQL Server 2005 , let DTA (Tunning
Advisor) to reccomed you what indexes are missed.
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel|||Joe,
many things just have said by others like replications, keep only used index,
etc, but I've said one more: FILLFACTOR.
If you have a very busy OLTP database, you can controll lock at least to
minimun with appropriate fillfactor.
I administrate many OLTP servers anda database and after read many articles
and try to find a baseline to controll lock. Finally, I've used these
templates with excelent result : Recreate every clustered index with
fillfactor 80% and recreate every nonclustered index with 90%.
You can increase or decrease this value that depends if you have updated
tables that need more, but stay with a little less in clustered than
noclustered.
Try this !
Kris
Joel Lyons wrote:
>It seems that a common problem we have is deadlocks occurring when one
>stored proc is updating a row and another store proc is running a read query
>against the same table. They'll often get deadlocked on two of the indexes
>for the table.
>For example:
>SP1 is updating a row in TableA, causing it to get an exclusive lock on
>TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
>SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
>attempts to get a shared lock on TableAIndex1.
>Deadlock!
>Of course, I can fix this by lowering the isolation level in SP2 to "read
>uncommitted", but I'd rather not.
>I can normally fix deadlocks by changing the order in which locks are
>acquired, but I don't know how to influence the order in which locks on
>indexes are acquired. Is it controlled by the order in which the fields are
>listed in a select or an update statement? If I standardize the order in
>which individual fields are listed whenever I perform a select or update
>from a given table, will that effect thr order in which index locks are
>acquired?
>Thoughts?
>Joel
--
Message posted via SQLMonster.com
http://www.sqlmonster.com/Uwe/Forums.aspx/sql-server/200804/1|||Wow! Those are some great ideas! I'll look into them further. Thank you.
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel|||Here is, IMHO, the bible for deadlock troubleshooting:
Kevin G. Boles
Indicium Resources, Inc.
SQL Server MVP
kgboles a earthlink dt net
"Joel Lyons" <JoelL@.novarad.net> wrote in message
news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
> It seems that a common problem we have is deadlocks occurring when one
> stored proc is updating a row and another store proc is running a read
> query against the same table. They'll often get deadlocked on two of the
> indexes for the table.
> For example:
> SP1 is updating a row in TableA, causing it to get an exclusive lock on
> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2 and
> attempts to get a shared lock on TableAIndex1.
> Deadlock!
> Of course, I can fix this by lowering the isolation level in SP2 to "read
> uncommitted", but I'd rather not.
> I can normally fix deadlocks by changing the order in which locks are
> acquired, but I don't know how to influence the order in which locks on
> indexes are acquired. Is it controlled by the order in which the fields
> are listed in a select or an update statement? If I standardize the order
> in which individual fields are listed whenever I perform a select or
> update from a given table, will that effect thr order in which index locks
> are acquired?
> Thoughts?
> Joel|||slip of the finger. here is the link:
http://blogs.msdn.com/bartd/archive/2006/09/09/Deadlock-Troubleshooting_2C00_-Part-1.aspx
--
Kevin G. Boles
Indicium Resources, Inc.
SQL Server MVP
kgboles a earthlink dt net
"TheSQLGuru" <kgboles@.earthlink.net> wrote in message
news:8aSdnTZNUZEe02vanZ2dnUVZ_jWdnZ2d@.earthlink.com...
> Here is, IMHO, the bible for deadlock troubleshooting:
>
> --
> Kevin G. Boles
> Indicium Resources, Inc.
> SQL Server MVP
> kgboles a earthlink dt net
>
> "Joel Lyons" <JoelL@.novarad.net> wrote in message
> news:4E93C8A1-6EAC-4522-9CB9-1CECEC9687C3@.microsoft.com...
>> It seems that a common problem we have is deadlocks occurring when one
>> stored proc is updating a row and another store proc is running a read
>> query against the same table. They'll often get deadlocked on two of the
>> indexes for the table.
>> For example:
>> SP1 is updating a row in TableA, causing it to get an exclusive lock on
>> TableAIndex1 and attempts to get an exclusive lock on TableAIndex2.
>> SP2 is reading TableA, causing it to get a shared lock on TableAIndex2
>> and attempts to get a shared lock on TableAIndex1.
>> Deadlock!
>> Of course, I can fix this by lowering the isolation level in SP2 to "read
>> uncommitted", but I'd rather not.
>> I can normally fix deadlocks by changing the order in which locks are
>> acquired, but I don't know how to influence the order in which locks on
>> indexes are acquired. Is it controlled by the order in which the fields
>> are listed in a select or an update statement? If I standardize the
>> order in which individual fields are listed whenever I perform a select
>> or update from a given table, will that effect thr order in which index
>> locks are acquired?
>> Thoughts?
>> Joel
>

Deadlock between Distribution Agent and Distribution Agent Cleanup

This is occurring regularly on SQL Server 2000 build 878.
The problem is a deadlock in the Distribution database. The Distribution
Agent spid is executing the SELECT statement below:
select @.max_xact_seqno = max(xact_seqno) from MSrepl_commands (READPAST)
where
publisher_database_id = @.publisher_database_id and
command_id = 1 and
type <> -2147483611
which is found in sp_MSget_repl_commands. It holds an Intent Shared page
lock on a data page in the MSrepl_commands table.
The Distribution Agent Cleanup spid is found to be running the command below:
DELETE MSrepl_commands WITH (PAGLOCK) where
publisher_database_id = @.publisher_database_id and
xact_seqno <= @.max_xact_seqno
located in the stored procedure sp_MSdelete_publisherdb_trans. This spid
holds an exclusive page lock on another data page in MSrepl_commands.
Both spids then attempt to obtain the same lock type on the page which is
locked by the other.
The Distribution Agent runs continuously and the Cleanup job is scheduled
for every 10 minutes. The Publication, Distribution and Subscription
databases are all on the same instance (3rd party vendor solution, not mine!)
in an active/active Win2003 cluster configuration. The articles are all
stored procedure executions.
Has anybody else seen this deadlock? Is it just a timing issue? Why is the
PAGLOCK hint used in sp_MSdelete_publisherdb_trans as above?
(I can't find any articles which correlate exactly to this problem)
Kind Regards
Andrew Pike
SQL Server DBA
Accenture UK
Do you have anonymous subscribers or named. With named subscribers the
distribution clean up agent cleans up more aggressively and you may see
problems like this when a subscriber has been offline for some time.
First off issue a select * from distribution.dbo.MSdistribution_status to
see how many undelivered vs delivered commands there are. If there are a
high number of delivered commands, I would stop the SQL Server Agent and run
the distribution clean up agent manually.
I can't comment on why the decision was made to implement the two types of
locks, but in general MS has done a lot of research to deliver optimal
performance. For example the 27 in sp_MSadd_repl_commands27 comes from
tests that they did to find the optimal number of commands to send to the
distribution database in a batch from the log reader agent. And yes, they
tested a range of commands to find which offered best performance.
It looks like the readpast is to prevent locking, and the page lock is to
prevent a table lock.
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
"Andrew Pike" <AndrewPike@.discussions.microsoft.com> wrote in message
news:CFB322D7-1D17-4064-AB36-6338A49C90A4@.microsoft.com...
> This is occurring regularly on SQL Server 2000 build 878.
> The problem is a deadlock in the Distribution database. The Distribution
> Agent spid is executing the SELECT statement below:
> select @.max_xact_seqno = max(xact_seqno) from MSrepl_commands (READPAST)
> where
> publisher_database_id = @.publisher_database_id and
> command_id = 1 and
> type <> -2147483611
> which is found in sp_MSget_repl_commands. It holds an Intent Shared page
> lock on a data page in the MSrepl_commands table.
> The Distribution Agent Cleanup spid is found to be running the command
> below:
> DELETE MSrepl_commands WITH (PAGLOCK) where
> publisher_database_id = @.publisher_database_id and
> xact_seqno <= @.max_xact_seqno
> located in the stored procedure sp_MSdelete_publisherdb_trans. This spid
> holds an exclusive page lock on another data page in MSrepl_commands.
> Both spids then attempt to obtain the same lock type on the page which is
> locked by the other.
> The Distribution Agent runs continuously and the Cleanup job is scheduled
> for every 10 minutes. The Publication, Distribution and Subscription
> databases are all on the same instance (3rd party vendor solution, not
> mine!)
> in an active/active Win2003 cluster configuration. The articles are all
> stored procedure executions.
> Has anybody else seen this deadlock? Is it just a timing issue? Why is
> the
> PAGLOCK hint used in sp_MSdelete_publisherdb_trans as above?
> (I can't find any articles which correlate exactly to this problem)
> Kind Regards
> Andrew Pike
> --
> SQL Server DBA
> Accenture UK
>