NetApp Insight 2015 – HOL (Hands on Lab)

Hi All,

This is neto from Brazil

How are you?

Please reserve your seat at NetApp Insight 2015 – HOL (Hands on Lab).

I already have one ­čśë

Thanks,

All the best,

neto
NetApp – I love this company!

Posted in Uncategorized | Tagged | Leave a comment

Linux Control Groups (cgroups) and Oracle 11.2.0.3

Hi All,

This is neto from Brazil

How are you?

Doing a POC and I had to use Linux Control Groups but … Oracle version is 11.2.0.3

4

But it works with 11.2.0.3 ­čÖé

init.ora =┬á*.processor_group_name=’oracle’

1

Running HammerOra with 120 users (48 virtual CPUs)

2

Changing Linux Control Groups on the fly:

echo 0-1 > /cgroup/cpuset/oracle/cpuset.cpus (for 2 Virtual CPUs)

3

All the best,

neto
NetApp – I love this company!

Posted in NetApp, netofrombrazil, Performance, workload tools | Tagged , , , , | Leave a comment

Capturing and Replaying Block Traces – blktrace, btrecord, btreplay and fio

Hi All,

This is neto from Brazil

How are you?

Here you have a quick tutorial how to take and replay block traces.

1) Check for entries in /sys/kernel/debug. If it is empty, please do:

mount -t debugfs nodev /sys/kernel/debug

2) Generate some IO:

dd if=/dev/dm-13 of=/dev/null bs=65536 iflag=direct

3) Capturing the trace:

cd traces
blktrace /dev/dm-13

4) From another windows, please send kill -15 to blktrace PID

We can use either btreplay or fio to replay the traces. Be careful, that fio replay test does writes and it is destructive to the target device.

Example with btreplay:

1) Convert to replay files:

cd  traces
btrecord -F

2) Replay the trace file

cd traces
btreplay -F -v -N /dev/dm-13

If you want to remap to a different disk device, create a file called newdisk.txt (e.g. dm-13 dm-15) and use:

cd traces
btreplay -M newdisk.txt -F -v -N

Example with fio:

1) Converting traces to blkparse format:

blkparse -i /root/traces/dm-13.blktrace.0 -o /dev/null -d /root/traces/dm-13-fio-trace

2) Using fio to replay:

fio –name=replay –filename=/dev/dm-13 –direct=1 –read_iolog=/root/traces/dm-13-fio-trace –replay_no_stall=1 –replay_redirect=/dev/dm-13

I will keep doing more tests and will post more updates soon.

Thank you very much for reading.

All the best

neto
NetApp – I love this company!

Posted in NetApp, Performance, workload tools | Tagged , , | 3 Comments

UPDATEv2 (easier): Automating database recovery using Pipes (Linux) with sqlplus – dynamic recovery (with scripts)

Hi All,

This is neto from Brazil

How are you?

Rodrigo found an easier way to do it…

I was talking with my good friend Leighton Nelson (https://twitter.com/@leight0nn) about how to create scripts to automate the auto recovery using pipes.

Captain Nascimento (https://twitter.com/@logwriter) and neto from Brazil (https:// twitter.com/@netofrombrazil) have developed some sample scripts to demonstrate how to use it.

Just to remember… AUTO for Oracle RAC database recovery does not work well, and we will need to execute some queries to find out the exactly change for the recovery.

select name, thread#, sequence#, first_change#, next_change# from v$archived_log where thread#=1 and 7728330471997 between first_change# and next_change#;

With the exactly change number, we will pass it as the STDIN for the sqlplus process (dynamically).

Procedure:

Please execute in the following order:

1) pipev2.sh: As Oracle user run: ./pipev2.sh &

#!/bin/bash

# ┬ápipev2.sh — Creating pipes for Oracle sqlplus (easy way) #
#  version 2.0
# ┬áby neto from Brazil – @netofrombrazil
# ┬áRodrigo Nascimento – @logwriter

rm -f /home/oracle/output
rm -f /home/oracle/sql
mkfifo /home/oracle/sql

cat > /home/oracle/sql <(sqlplus / as sysdba < /home/oracle/sql > /home/oracle/output)

NOTE: sql is the pipe for STDIN and output is for STDOUT

Terminal 2: As Oracle user run:

echo “select * from v\$instance;” > /home/oracle/sql

You can check the output doing: tail ÔÇôf /home/oracle/output
NOTE: Sending “select * from v$instance” to pipe sql – where it will be redirected to STDIN for sqlplus process

Now, it will be easier to execute queries to find out the change number (check STDOUT) and create dynamic queries to send to sqlplus process (STDIN).

echo “recover database using backup controlfile until cancel;” > /home/oracle/sql
echo “+DG_ARCH/prod/archivelog/2014_01_06/thread_1_seq_27236.1832.836139501” > / home/oracle/sql

Thank you very much for reading.

Thanks my brother Rodrigo.

All the best,

neto
NetApp – I love this company!

Posted in Uncategorized | Leave a comment

11 (in binary) nice pictures from NetApp Insight 2014 in Berlin

Hi All,

This is neto from Brazil

How are you?

I would like to share 11 (in binary) nice pictures from NetApp Insight 2014 in Berlin.

This is me:

neto - NetApp - Osoq - NetApp Logo - black shoes - NetApp Communities - 350X263 - transparency

This is also me ­čÖé

IMG_0467

Another nice one:

IMG_0480

NetApp – I love this company!

All the best,

neto
NetApp – Culture

Posted in #NTAPInsight, NetApp, netofrombrazil | Tagged , , | Leave a comment

Performance sample and Blessed Thanksgiving to all

Hi All,

This is neto from Brazil

How are you?

This is my first post on my new blog.

Here, I am going to talk about NetApp, Databases (Oracle) and Performance.

I hope you guys enjoy.

Here is a very simple sample about Total IOPS and Average Latency

Snip20141126_5

Thank you very much for reading.

I wish to all a blessed Thanksgiving.

All the best,

neto
NetApp – I love this company!

Posted in Performance | Tagged | Leave a comment