Detailed Report all zero

Bug #1181576 reported by Markus Falb
26
This bug affects 6 people
Affects Status Importance Assigned to Milestone
MOVED - Percona Playback
Confirmed
Undecided
Unassigned

Bug Description

I run with --input-plugin tcpdump
The report shows real numbers but the numbers in the Detailed Report are all zero

...
Error Report finished

Detailed Report
----------------
SELECTs : 0 queries (0 faster, 0 slower)
INSERTs : 0 queries (0 faster, 0 slower)
UPDATEs : 0 queries (0 faster, 0 slower)
DELETEs : 0 queries (0 faster, 0 slower)
REPLACEs : 0 queries (0 faster, 0 slower)
DROPs : 0 queries (0 faster, 0 slower)

Report
------
Executed 76 queries
Spent 00:00:00.222713 executing queries versus an expected 00:00:00.014925 time.
5 queries were quicker than expected, 71 were slower
A total of 0 queries had errors.
Expected 27 rows, got 27 (a difference of 0)
Number of queries where number of rows differed: 0.

Average of 76.00 queries per connection (1 connections).

Tags: i49245
Revision history for this message
Przemek (pmalkowski) wrote :

Happens also when slow log is used:

percona-playback --dispatcher-plugin=thread-pool --thread-pool-threads-count=24 --mysql-schema=test --query-log-file=slow1.log
(...)
Detailed Report
----------------
SELECTs : 0 queries (0 faster, 0 slower)
INSERTs : 0 queries (0 faster, 0 slower)
UPDATEs : 0 queries (0 faster, 0 slower)
DELETEs : 0 queries (0 faster, 0 slower)
REPLACEs : 0 queries (0 faster, 0 slower)
DROPs : 0 queries (0 faster, 0 slower)

Report
------
Executed 32 queries
Spent 00:00:12.128907 executing queries versus an expected 00:08:46.736992 time.
18 queries were quicker than expected, 13 were slower
A total of 1 queries had errors.
Expected 14 rows, got 113 (a difference of 99)
Number of queries where number of rows differed: 16.

Average of 3.56 queries per connection (9 connections).

[root@pxc2 ~]# percona-playback --version
percona-playback
Version: 0.7

Changed in percona-playback:
status: New → Confirmed
tags: added: i49245
Revision history for this message
Andrew Davidoff (davidoff) wrote :

With the caveat that I don't fully understand why the history here, I think this commit is what broke this, at least for me:

commit 523cb5d118cdbb0e0cb22564bd1b78fed4e5440c
Author: Frederic Descamps <email address hidden>
Date: Wed Oct 3 11:20:44 2012 +0200

    fix a bug for reporting

That commit adds a leading newline to the query search that's looking for the type of statement. Ex:

- if (new_query.find("SELECT ") != std::string::npos)
+ if (new_query.find("\nSELECT ") != std::string::npos)
     {
                nr_select++;
                nr_select_faster+= faster;
                nr_select_slower+= slower;
     }

I haven't looked at C++ in years but as far as I can tell, there's no reason there'd every be a leading newline here unless there was one in the slow log, and in my case, that doesn't seem to happen.

I am hesitant to say fixing this is as simple as removing the leading newlines in these find calls because there must be a good reason the commit was made in the first place, but I do not understand it at the moment.

Revision history for this message
Andrew Davidoff (davidoff) wrote :
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.