Hi all,
I have replication running on a sql server 2008 publisher/distributor, with a sql server 2005 as a subscriber.
It has been running fine for several weeks. I've been opening the log reader monitor often just to check on it and normally I never see anything but:
So that's fine. However, the other day someone said a bit of data was stale--someone I trust and if he says the data's old I will take note. So I looked at the logreader agent monitor and in addition to the normal message, there were two lines that were nearly identical that said:
The timestamp of these lines coincided with around the same time the report with stale data was run against the replicated database. Yes it says zero records marked but why all of a sudden tell me this when it never told me that before?
So I've been trying to figure out what that message means. And, if it's "normal", why is it just showing up now after nearly a month? If it's not normal, what does it indicate is wrong?
When I looked at the data that was alleged to be stale it had updated by then so I couldn't prove anything. But I think it's no coincidence that those 'abnormal' messages showed up exactly when people who have eagle-eyes on this data tell me they saw stale data.
Other info: We are in Full recovery mode, and backup/truncate the log every hour. As another facet to this question--I wonder if it's possible that the log reader didn't fully replicate everything prior to the backup--ie, maybe a huge bulk update occured 5 seconds before the backup--then before replication could complete the backup pulled un-replicated records out from under it by truncating the log?
Thanks for any insight on this.
--Jim
I have replication running on a sql server 2008 publisher/distributor, with a sql server 2005 as a subscriber.
It has been running fine for several weeks. I've been opening the log reader monitor often just to check on it and normally I never see anything but:
So that's fine. However, the other day someone said a bit of data was stale--someone I trust and if he says the data's old I will take note. So I looked at the logreader agent monitor and in addition to the normal message, there were two lines that were nearly identical that said:
The timestamp of these lines coincided with around the same time the report with stale data was run against the replicated database. Yes it says zero records marked but why all of a sudden tell me this when it never told me that before?
So I've been trying to figure out what that message means. And, if it's "normal", why is it just showing up now after nearly a month? If it's not normal, what does it indicate is wrong?
When I looked at the data that was alleged to be stale it had updated by then so I couldn't prove anything. But I think it's no coincidence that those 'abnormal' messages showed up exactly when people who have eagle-eyes on this data tell me they saw stale data.
Other info: We are in Full recovery mode, and backup/truncate the log every hour. As another facet to this question--I wonder if it's possible that the log reader didn't fully replicate everything prior to the backup--ie, maybe a huge bulk update occured 5 seconds before the backup--then before replication could complete the backup pulled un-replicated records out from under it by truncating the log?
Thanks for any insight on this.
--Jim
