pt-table-checksum skipping tables

[Amar]

Hi,

I am using pt-table-checksum for checking data consistency, the
problem I am facing is that its skipping the tables which got correct
rows and showing this output

05-31T19:58:16 Skipping table database.table because on the master it
would be checksummed in one chunk but on these replicas it has too
many rows:
26017 rows on slave1
27213 rows on slave2
27100 rows on slave3
The current chunk size limit is 22440 rows (chunk size=11220 * chunk
size limit=2.0).

If I do count on database.table it shows 22101 rows on all slaves
including master.

This is what I am running

pt-table-checksum h=localhost,P=3306,u=user,p=pass --nocheck-
replication-filters --lock-wait-timeout=120 --recursion-method
dsn=localhost,D=percona,t=dsns --ignore-tables
table1,table2,table3,table4 --databases-regex ^database

Any help is appreciated.

Thanks,

[Baron Schwartz]

Amar,

There are some previous threads on this group that are probably related
to your topic.

- Baron

--
Chief Performance Architect at Percona <http://www.percona.com/>
+1 (888) 401-3401 x507
Calendar: <https://tungle.me/percona_baron>
Percona Live New York, Oct 1-2 <http://www.percona.com/live>

[Amar]

Thanks Baron. I have read previous posts which suggest of increasing --
chunk-size-limit and --chunk-time.

Its not really clear from post that both was used together or chunk-
time did the trick. And the version in that post is
2.0.3 and I am using 2.1.1

Please suggest.

Thanks,
Amar