question

robvolk avatar image
robvolk asked

Trace Replay Problems

A little background: I need to record a trace in production and replay in another environment. Production has sensitive data that must be scrubbed from the trace before it can be moved out of that environment. I've imported the trace into a SQL table, scrubbed the data, and exported for playback to a new trace file. I've also bcp'd into a new trace table. Whenever I try to replay the scrubbed trace, it does nothing. The replay stops immediately and I get the following output:

Replay server: servername
Total events = 595354
Replayed events = 0
Provider errors = 0
Internal errors = 0
Total errors = 0
Provider hit ratio = 0.00%
Internal hit ratio = 0.00%
Overall hit ratio = 0.00%
Replay time = 0 hr 0 min 0 sec 250 ms
Per Books Online I've made sure logins match on both servers. The databases also match by name, but not by ID. I've tried replaying both the scrubbed file and table, but get the same results. I'm about to replay the raw trace on another production machine just to see if it works at all, but ultimately I need a scrubbed trace that can replay elsewhere. Has anyone done something similar? Or can point out problems I might have missed? Has anyone modified trace data and replayed it successfully? Thanks. UPDATE: the trace played back successfully with a few minor errors. I will next attempt a playback with a truncated portion of the trace.
profilertracereplay
10 |1200

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

1 Answer

·
ShakyKnees avatar image
ShakyKnees answered
I tried three times unsuccesfully to alter the DBID values in the trace file by exporting from Profiler to a SQL table, updating, and re-importing the table to Profiler as trace data. Profiler refused to Replay it, but I forgot what it complained about. However, I found it easy to renumber the DBID values of the target databases in my sandbox by DETACHing and reATTACHing them. This made it possible to replay my production trace in the sandbox. I have not found a credible explanation of the "hit ratio" report. I've seen a guess that it is about cache hit ratios, but I doubt it. ,I tried three times unsuccesfully to alter the DBID values in the trace file by exporting from Profiler to a SQL table, updating, and re-importing the table to Profiler as trace data. Profiler refused to Replay it, but I forgot what it complained about. However, I found it easy to renumber the DBID values of the target databases in my sandbox by DETACHing and reATTACHing them. This made it possible to replay my production trace in the sandbox. I have not found a credible explanation of the "hit ratio" report. I've seen a guess that it is about cache hit ratios, but I doubt it.
10 |1200

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Write an Answer

Hint: Notify or tag a user in this post by typing @username.

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.