pt-stalk doesn't collect the correct number of mysqladmin samples
Bug #1166414 reported by
Ryan Lowe
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| Percona Toolkit moved to https://jira.percona.com/projects/PT |
Triaged
|
Undecided
|
Unassigned | ||
Bug Description
When I run:
%> pt-stalk --no-stalk --iterations 1 --run-time=3600
...
It finishes correctly, but:
%> grep -c Bytes_sent /var/lib/
3592
%> grep -c TS /var/lib/
3600
Suggested behavior would be to have all tools collect the same number of iterations :)
| tags: | added: pt-stalk |
Revision history for this message
| Daniel Nichter (daniel-nichter) wrote | #1 |
| Changed in percona-toolkit: | |
| status: | New → Triaged |
| summary: |
- pt-stalk collections are not consistent + pt-stalk doesn't collect the correct number of mysqladmin samples |
Revision history for this message
| Shahriyar Rzayev (rzayev-sehriyar) wrote | #2 |
To post a comment you must log in.
mysqladmin is working on its own. So pt-stalk tells it to get 3600 samples (-i1 -c3600), but either that program is inexact or pt-stalk is killing it too soon. The other data like -df is inside a loop controlled by pt-stalk, so it's more or less guaranteed to get the number of samples. This loop should also allow mysqladmin to run for the full 3600 seconds, but I guess not. I suppose we could run mysqladmin in the loop too?