RE: Snapshot Help

Tech-Archive recommends: Fix windows errors by optimizing your registry

Hi JLS,

My apologies for the lack of response on this so far. Hilary was indeed
correct in saying that it is not possible to manually perform just the
distribution database portion of snapshot generation processing. To find out
if you can resolve the issue, you may want to first find out who else is
involved in the deadlock (the snapshot agent is obviously a participant)
using either DBCC TRACEON(1204,-1, 3605) or the SQL Profiler. Once you know
who else is involved in the deadlock, you may be able to "resolve" the
deadlock if you can consider doing one of the following:

1) Is it possible to schedule the processing of the other connections to run
while the snapshot is not being generated?
2) If the other connections involved are from a user application, is it
possible to rearrange its locking order to avoid deadlocking with the
snapshot agent?
3) Is it possible to lower the deadlock priority for the "other" connections
and then implement deadlock retry logic on them?

The real reason I am posting this is that we do make snapshot generation
processing more resilient to deadlocks in SQL2005 (deadlock retry logic,
configurable deadlock priority in the snapshot agent). So in the interest of
product improvement, it would be great if you can let me know the command
that the snapshot agent was deadlocking on. Also, it would be really helpful
if you can let me know whether you are using concurrent (Do not lock
table...) snapshot. I am guessing that you are from the symptoms that you
provide below but I want to make sure.

-Raymond

"JLS" wrote:

> Is there a way to have just the posting into the Distribution database portion of a snapshot resubmit after some process blocks the snapshot?
>
> This is the 3rd time the job is attempting to get the snapshot completed. After 50 minutes of processing it became deadlocked & was chosen as the victim.
>
> ANY ADVICE will be GREATLY appreciated. I can't get Replication backup until the Snapshot is in sync.
>
> Totally unrelated question.....
> Also, what table can I hack when I have an error & just want to delete those rows from attempting to be replicated?
>
> Thanx!
> JLS
.

Relevant Pages

  • Re: [PATCH 3/3] Add timeout feature
    ... do the snapshot op. ... run_frozen -t timeout mountpoint do-snapshot ... Userland apps can be swapped out and need kernel memory allocations ... even trivial applications can deadlock. ...
    (Linux-Kernel)
  • Re: Hangs with UFS2 snapshots
    ... > intensive and generates another snapshot). ... > beige box with a Mylex AcceleRAID controller, but I can reproduce the same ... can break to it when the deadlock occurs (see the developers' ...
    (comp.unix.bsd.freebsd.misc)
  • [git pull] 2.6.28-rc2 device-mapper fixes
    ... dm snapshot: fix register_snapshot deadlock ... wait for chunks in destructor ...
    (Linux-Kernel)
  • Re: Snapshots, Validation, Resynchronization with Merge Replicatio
    ... You should have your distribution database on your ... In your topology when you run the snapshot agent, ... What this means is that a subscriber has 14 days to sync at least ... NOw the question is how often do you have to run the snapshot. ...
    (microsoft.public.sqlserver.replication)
  • Re: Snapshot Creation
    ... Distributor/Publisher with my AD account (Windows administrator and SQL ... But making my AD account the owner of the Snapshot Agent job, ... Looking for a SQL Server replication book? ...
    (microsoft.public.sqlserver.replication)