process could not execute 'sp_replcmds'

Question

process could not execute 'sp_replcmds'

ashish.johri

Mr or Mrs. 500

Points: 560
More actions
December 8, 2008 at 4:19 am

#216177

I am getting following information again and again in the job log
agent DAF1-LM13-AshKpPartition13-17 failed. The process could not execute 'sp_replcmds' on 'DAF1-LM13'.;
Once this fails and then the agent starts working again.
I found in one google link that this can be avoided by increasing the timeout period or increasing the buffer size. I am not sure why this comes. But I feel this is a general error and many DBAs would be facing this. Any heads up!!
PS: 'The name of the agent and partition is changed intentionally'
Regards,
Ashish

Viewing 6 posts - 1 through 5 (of 5 total)

You must be logged in to reply to this topic. Login to reply

ashish.johri Mr or Mrs. 500 Points: 560 More actions · Answer 1

No replies :(.

What is the reason of getting "agent DAF1-LM13-AshKpPartition13-17 failed. The process could not execute 'sp_replcmds' on 'LLW1-DD2'"

Mike_D Hall of Fame Points: 3295 More actions · Answer 2

Is there no other clue in the error message? Normally, you'd expect to find a reason for the failure such as "general network error" or "query timeout" etc. Just as an example, you could have problems if you are log reading across a WAN and your transaction sizes are large. If that bit of the network is not performing well this is the sort of situation in which the log reader agent can struggle. For reasons that are specific to my environment, I like to keep my publishers' transactions to 1000 rows or less.

You can normally determine this by checking delivery rates and latency. But, from what you have said, it seems like your replication is working so I'd just ignore it. Some of my publishers are on rather dodgy bits of network so I have special log reader agent profiles to cater for this (increase QueryTimeout and decrease ReadBatchSize).

ashish.johri Mr or Mrs. 500 Points: 560 More actions · Answer 3

This message comes once in every 2 hours. I increased the query timeout period and could resolve this. I think that slow n\w connection causes less number of articles delivered.

-Ashish

abhij33t SSC Eights! Points: 899 More actions · Answer 4

I am having the same problem but, the reason it gives to me is strange enough. It says "can not find any user/group with the name 'dbo' ". I am using a domain level administrator account for all of this on a test vm setup. snapshot agent is running fine with the same security settings, it's just the log reader agent which doesn't likes it.

luckysql.kinda SSCertifiable Points: 7816 More actions · Answer 5

I think this was an issue with sql2k5 SP1. After that this error is no more. Could you tell your @@version Abhijeet?