July 19, 2013 at 5:35 am
We are currently in the process of migrating our environments. We have multiple publication servers, a central distributor, and multiple subscribers. We are migrating all parts of the environments.
The first step happened a couple of weeks ago, we migrated a publishing server to a new host and changed the distributor, replication was dropped and recreated without any problems.
The second step is just to change the distributor of another server, this consists of 4 transactional publications. Three went in with any problems the final one was created without issues, the snapshot generated, but the subscription failed with the following error:
The distribution agent failed to create temporary files in blah\blah\COM
I have given the distribution agent account access to the directory on the subscriber but it still fails as we have other servers working this way I do not understand why this is failing.
It think that this may be down to the qreader agent, we didn't use one of these for the first server, which I believe is used for transactional replication with updateable subscriptions which we are not using.
My next step, as this is now affecting production reporting, is to strip out the replication and start from scratch but this time without the qreader agent.
Any comments or recommendations would be appreciated.
July 20, 2013 at 11:49 am
20-July-2013
Taken out replication overnight, disabled qreader job on distributer.
Recreated all four publications - no problems all snapshots created.
Three subscriptions work fine.
The last one again fails because of the following error:
The distribution agent failed to create temporary files in 'c:\Program Files\Microsoft SQL Server\110\COM' directory. System returned errorcode 5. (Source: MSSQL_REPL, Error number: MSSQL_REPL21100)
This is not an updateable subscription, the snapshot is over 35GB which shouldn't be an issue. I believe that the c: drive referred to is that on the subscriber - which the distribution agent has full access to.
Any ideas?
July 23, 2013 at 1:30 am
Now resolved. The drive replication required access to was on the C: drive of the distributor.
I can't see why it would need it for this specific publication/subscription and not for the other 4 transactional and 4 snapshot publications using this distributor, unless it was down to size as the snapshot for this one was over 35GB.
This does add an extra problem as this is a virtual instance on a cluster so I now have to give the account access to the directory on the C: on all 4 cluster nodes.
any suggestions anyone?
Thanks
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply