Logreader failed to construct replicated command from LSN

I have this error coming up 'The process could not execute 'sp_replcmds' on
'VDB1' which always seems to accompany 'Logreader failed to construct
replicated command from LSN {0008e253:00009ddb:0003}.'
I turned on verbose logging and it gave this:
2007-05-04 01:31:16.018 Publisher: {call sp_replcmds (500, 0, 0, , 15,
500000)}
2007-05-04 01:31:16.065 I~0x0008e25300009ddb0005~1:
Looking up that sequence number with:
exec sp_browsereplcmds @.xact_seqno_start = '0x0008e25300009ddb0005'
, @.xact_seqno_end = '0x0008e25300009ddb0005'
Returned nothing.
I also tried:
select * from MSrepl_transactions
where xact_seqno = '0x0008e25300009ddb0005'
It returned nothing also.
This same error is preventing 2 publications, but other 8 are working fine.
I have stopped and started log reader job a couple of times.
Any ideas?
Thanks,
Chuck Lathrope
www.sqlwebpedia.com
I've bumped into this from time to time. What you need to do is to run
profiler with the xactSequence number column.
Next time you get this error, look up the offending statement, for example
it could be a proc which is writing something unexpected to the database.
Last time I got this, it was an update to a text column with a ''.
"Chuck Lathrope" <computerguy_chuck@.fixmehotmail.com> wrote in message
news:5likshd2rlr9.1cei74wrqkxlw$.dlg@.40tude.net...
>I have this error coming up 'The process could not execute 'sp_replcmds' on
> 'VDB1' which always seems to accompany 'Logreader failed to construct
> replicated command from LSN {0008e253:00009ddb:0003}.'
> I turned on verbose logging and it gave this:
> 2007-05-04 01:31:16.018 Publisher: {call sp_replcmds (500, 0, 0, , 15,
> 500000)}
> 2007-05-04 01:31:16.065 I~0x0008e25300009ddb0005~1:
> Looking up that sequence number with:
> exec sp_browsereplcmds @.xact_seqno_start = '0x0008e25300009ddb0005'
> , @.xact_seqno_end = '0x0008e25300009ddb0005'
> Returned nothing.
> I also tried:
> select * from MSrepl_transactions
> where xact_seqno = '0x0008e25300009ddb0005'
> It returned nothing also.
> This same error is preventing 2 publications, but other 8 are working
> fine.
> I have stopped and started log reader job a couple of times.
> Any ideas?
> Thanks,
> Chuck Lathrope
> www.sqlwebpedia.com
|||On Fri, 4 May 2007 08:29:28 -0400, Hilary Cotter wrote:
Thanks, it only seems to be getting worse for me. The replication monitor
can't expand the server nodes now. I restarted sql service on the
distributor and that didn't help. Has anyone seen that?
I am setting the -ReadBatchsize to 1 as Jonathan suggested, hopefully it
can find to correct issue as this command in error reports looks to be
pointed to something already completed.
-Chuck
[vbcol=seagreen]
> I've bumped into this from time to time. What you need to do is to run
> profiler with the xactSequence number column.
> Next time you get this error, look up the offending statement, for example
> it could be a proc which is writing something unexpected to the database.
> Last time I got this, it was an update to a text column with a ''.
>
> "Chuck Lathrope" <computerguy_chuck@.fixmehotmail.com> wrote in message
> news:5likshd2rlr9.1cei74wrqkxlw$.dlg@.40tude.net...
|||Update:
One of the two publications started to work, not sure why or when.
The batchsize to 1 still brought back the same LSN error - which is still
returning nothing when I query it with:
select * from MSrepl_transactions
where xact_seqno = '0x0008e25300009ddb0005'
We are going to kill the publications and recreate them - luckily they are
not super critical like the others.
Still having Replication Monitor issues (Still running SP1 on server; SP2
on my local box).
-Chuck
On Fri, 4 May 2007 08:29:28 -0400, Hilary Cotter wrote:
Thanks, it only seems to be getting worse for me. The replication monitor
can't expand the server nodes now. I restarted sql service on the
distributor and that didn't help. Has anyone seen that?
I am setting the -ReadBatchsize to 1 as Jonathan suggested, hopefully it
can find to correct issue as this command in error reports looks to be
pointed to something already completed.
-Chuck
[vbcol=seagreen]
> I've bumped into this from time to time. What you need to do is to run
> profiler with the xactSequence number column.
> Next time you get this error, look up the offending statement, for example
> it could be a proc which is writing something unexpected to the database.
> Last time I got this, it was an update to a text column with a ''.
>
> "Chuck Lathrope" <computerguy_chuck@.fixmehotmail.com> wrote in message
> news:5likshd2rlr9.1cei74wrqkxlw$.dlg@.40tude.net...
|||While recreating the publications, the snapshot errored because of an
obsolete XSD for an XML table. We have the option set to false to publish
the XSD's, so not sure how this came about.
Now that these are working, the Replication Monitor server nodes can be
expanded now.
-Chuck
On Fri, 4 May 2007 10:15:26 -0700, Chuck Lathrope wrote:
[vbcol=seagreen]
> Update:
> One of the two publications started to work, not sure why or when.
> The batchsize to 1 still brought back the same LSN error - which is still
> returning nothing when I query it with:
> select * from MSrepl_transactions
> where xact_seqno = '0x0008e25300009ddb0005'
> We are going to kill the publications and recreate them - luckily they are
> not super critical like the others.
> Still having Replication Monitor issues (Still running SP1 on server; SP2
> on my local box).
> -Chuck
>
> On Fri, 4 May 2007 08:29:28 -0400, Hilary Cotter wrote:
> Thanks, it only seems to be getting worse for me. The replication monitor
> can't expand the server nodes now. I restarted sql service on the
> distributor and that didn't help. Has anyone seen that?
> I am setting the -ReadBatchsize to 1 as Jonathan suggested, hopefully it
> can find to correct issue as this command in error reports looks to be
> pointed to something already completed.
> -Chuck