Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
niravbhat
Starting Member
2 Posts |
 Posted - 2011-09-01 : 12:49:18
|
| We finally get them to purge their data. The issue you had on the weekend of the 6th was that they put in a -1 in the parameter, then finally realized they made a mistake and put a value in the parameter on the 4th. Then the program wouldn’t run because of conflicts, when it finally ran, I was able to shrink the files. That was the 8th ,.Then we had the wandering 50GB, and for some reason their BO reports stopped working. On the 10th according to the logs – replication started throwing errors.I get asked about it on the 16th, I do the check of the replication monitor and it says everything is green .When they say they have an issue, I reinitialized their publication and made sure all the active agent jobs are running (3 were not because of login id), they don’t get back to me till the next day.On the 18th I finally reached a point where I shut down everything and restarted the replication process using the two sql queriesThat opened up the distribution database, which had reached 46GB, with records that the subscriptions were not picking up. That in turn grew the transaction log which created a space issue on the E drive where we keep the distribution database. Once the 46 Gb of transactions was processed, I was able to shrink the log and the database to their normal size.I ran a query and there is a lack of data between 8.10 and 8.19.2011. The system is supposed to keep itself in-sync, but apparently it has processing delays.the replication monitor is useless as far as determining if there is an error. Other than a gross problem, such as the agent being stopped, it shows nothing.So the only solution is to stop the process and then restart it. |
|
|
russell
Pyro-ma-ni-yak
5072 Posts |
Posted - 2011-09-01 : 12:58:03
|
| If you need to pick up the missing data, you'll need to run the snapshot agent.Also, seems you need to extend the expiration period, and retention period. However, you need to actively monitor transaction log size and distribution db and log size.You should create custom alerts to notify you before it's too late.Easy to create a job to drop tracer tokens and measure latency too, then page you if it's too high.No one noticing for 6 days is pretty shocking! You really need to create some monitoring jobs. Also, have a look at sp_publication_validation |
 |
|
|
|
|
|
|
|
replication issues