Data Virtualizer

Amazon RDS Performance Insights (PI) doesn't have a single pane of glass dashboard for clusters, yet. Currently PI has a dashboard that has to be looked at for each instance in a cluster. On the other hand one can create a simple cluster dashboard using Cloudwatch. PI, when enabled, automatically sends three metrics to Cloudwatch every minute. These metrics are DBLoad DBLoadCPU DBLoadNonCPU DBLoad = DBLoadCPU + DBLoadNonCPU These metrics are measured in units of Average Active Sessions (AAS). AAS is like the run queue on the UNIX top command except at the database level. AAS is the average number of SQL queries running concurrently in the database. In the ...

khailey
Uncategorized

“delayed commit ok initiated" -  is a thread state in Aurora MySQL which indicates the thread has started the async commit process but is waiting for it to be ack’d. You will not find this thread state in MySQL as  MySQL  does not use our async commit protocal, it is Aurora MySQL specific. This is  usually the genuine commit time of a transaction. This is a "state" and not a wait.

khailey
Uncategorized

Installing CLI on LINUX 1. install PIP https://docs.aws.amazon.com/cli/latest/userguide/awscli-install-linux.html#awscli-install-linux-pip curl -O https://bootstrap.pypa.io/get-pip.py python get-pip.py --user 2. install AWS CLI https://docs.aws.amazon.com/cli/latest/userguide/installing.html pip install awscli --upgrade --user 3. configure aws configure For “aws configure” you will need AWS Access Key ID: AWS Secret Access Key: Which you can get by going to the AWS console, going to IMS and creating access key. Running example Once “aws” is configured you can run the CLI like aws \ pi get-resource-metrics \ --region us-east-1 \ --service-type RDS \ --identifier db-xxxxxx \ --metric-queries "{\"Metric\": \"db.load.avg\"}" \ --start-time `expr \`date +%s\` - 60480000 ` \ --end-time `date +%s` \ --period-in-seconds 86400 That “—identifier” is for one of my databases, so you will have to change ...

khailey
Uncategorized

Thanks to Jeremiah Wilton for the following info: This wait event indicates that there is a thread which is waiting on an InnoDB record lock. Check your database for conflicting workloads. More information on InnoDB locking can be found here: https://dev.mysql.com/doc/refman/5.6/en/innodb-locking.html   In other words, record-level lock conflicts are happening. More than one connection is trying to update the last_login for a particular id in the_table at the same time. Those connections are conflicting and serializing on the record lock for that id. Here's a query that can help you identify the blocker and waiter for InnoDB record locks in MySQL-family engines. Run ...

khailey
Uncategorized

If you think so, check out shell script "ioh.sh" from github at  https://github.com/khailey/ioh Introduction and Goals The goal of ioh.sh is to measure both the throughput and latency of the different code layers when using NFS mounts on a ZFS appliance. The ZFS appliance code layers inspected with the script are I/O from the disks, ZFS layer and the NFS layer. For each of these layers the script measures the throughput, latency and average I/O size. Some of the layers are further broken down into other layers. For example NFS writes are broken down into data sync, file sync and non-sync operations and NFS ...

khailey
Uncategorized

I always forget IP address range coverage rules and forget where to look. It's the wiki! https://en.wikipedia.org/wiki/Classless_Inter-Domain_Routing and for good reference here is the table: /32 is for a single address /24 is for a range in the last place x.x.x.0 /16 is for a range in the last 2 places x.x.0.0

khailey
Uncategorized

How do you determine if the Oracle SQL optimizer has created a sub-optimal execution plan? re-run statistics and see what happens? wait for Oracle to find other execution plans? What if neither method is helping? Do you read the execution plan? What do you look at? Differences in actual vs estimated? How successful is that? Look for full table scans?  Do you look at the 10053 trace? How much time and effort does that take?  What do you look at in the 10053 trace. Do you have a systematic methodology  that works in almost all cases? Well there is a method that is reliable ...

khailey
Uncategorized

  http://www.oaktable.net/blog/oak-table-world-2017-oracle-open-world The Oak Table members will be discussing their latest technical obsessions and research on Monday and Tuesday, (Oct. 2nd and 3rd, 2017).  The truth is, folks-  The Oak Table experts are an AWESOME group, (if I don’t say so myself! :)) as we could have easily done another day of incredible sessions, but alas, two days is all we have available for this year’s event.  

khailey
Uncategorized

  While at Delphix, we did a lot of storage benchmarking. The I/O response times of Delphix depends, as one would logically imagine, heavily on the underlying disks. Sure Delphix can cache a lot ( with 1 TB of ram and 3x compression that's 3TB and that 3TB can be shared by 10 or a 100 copies being the equivalent to 30TB or 300TB of databases) but really there will always be important I/O coming from the storage subsystem. Now Delphix mainly runs databases loads, so the best test for storage that is hooked up to Delphix is to benchmark the storage I/O ...

khailey
Uncategorized

' photo by Thomas Hawk https://github.com/khailey/pgmon $ ./pgmon.sh Usage: pgmon.sh [username] [password] [host] <sid=postgres> <port=5432> <runtime=3600> $ ./pgmon.sh kyle kyle mymachine.com psql -t -h mymachine.com -p 5432 -U kyle postgres < /tmp/MONITOR/tmp/mymachine.com:postgres_collect.pipe & RUN_TIME=-1 COLLECT_LIST= FAST_SAMPLE=wts TARGET=mymachine.com:postgres DEBUG=0 Connected, starting collect at Wed Mar 8 12:05:12 PST 2017 starting stats collecting SET AAS| blks_hit | blks_read | tup_returned | tup_fetched | tup_inserted | tup_updated | tup_deleted 1 | 38281 | 1 | 628629 ...

khailey
Uncategorized