Viewing 15 posts - 1 through 15 (of 24 total)
is there a pattern or some factor in common among the non-replicated rows?
September 23, 2004 at 3:06 pm
i'd go with merge . . . it's the most autonomous and best at synchronization. but, beware the complexities of this most elegant of replication models. it's for well designed...
September 22, 2004 at 9:00 am
excellent article . . . both for the "gotcha" and as a reminder to do your homework . . . lookup "Using Data Types" in bol. type AND length are specified as...
July 15, 2004 at 7:47 am
the CHECKSUM() function looks like a good candidate to investigate.
where CHECKSUM(Col) = CHECKSUM(Val)
June 3, 2004 at 8:42 am
if IMPLICIT_TRANSACTIONS is on, then the explicit tran plus the implicit tran started by the first update would make the trancount 2. by using the optional "TRANSACTION" keyword with "ROLLBACK", the...
May 19, 2004 at 2:58 pm
looks like you told it to reinitialize, but didn't tell it to do a new snapshot before hand . . . @force_invalidate_snapshot = 1
now that the deed is done:
May 17, 2004 at 12:07 pm
pattern matching (PATINDEX or LIKE) is required . . . when you want "4", for example, a value of "14" would match also with CHARINDEX. mssql doesn't have the strongest regular expressions support. ...
May 17, 2004 at 11:08 am
all that's in the bol referrence . . .
"Rename the warm standby server to the name of the original Publisher. Replication will continue to distribute data changes to Subscribers."
May 15, 2004 at 10:19 am
the original issue was removing leading tabs. nulls and blanks have none. hence, like any other string without leading tabs, they are simply returned...
May 15, 2004 at 10:09 am
xp_sendmail opens its own connection . . . hence, local temp tables may be out of scope. you might want to use a global temp table and the proc's "@dbuse"...
May 14, 2004 at 7:21 am
just the obvious . . . you're upping the potential impact on replication latency, network usage, differing datatypes (don't want to send a 3000 character text entry to a nvarchar(2000) column) ....
May 12, 2004 at 2:01 pm
it's possible . . . see bol "Strategies for Backing Up and Restoring Transactional Replication"
May 12, 2004 at 12:47 pm
only two ways i know . . . both return a query result set:
sql-dmo - "EnumJobInfo" method
stored proc - "sp_helpmergepullsubscription"
both are in bol.
May 10, 2004 at 12:41 pm
was an initial snapshot of the publication generated (yes, even for merge replication)?
May 10, 2004 at 11:59 am
Viewing 15 posts - 1 through 15 (of 24 total)