Skip to content

Latest commit

 

History

History
94 lines (89 loc) · 3.31 KB

DAY19.md

File metadata and controls

94 lines (89 loc) · 3.31 KB
Raw

DAY 19

Redhat Traininig

Stratis

  • It works on thin provision
  • It allocates memory on run time
  • 2**24 filesystem can be created from single Pool
  • it formats the filesystem as well
  • dm-raid - > handles the system output in terms of storage
  • maximum size of Pool can be only 1 terabyte
  • Block-Devices refers to the hard disk
  • By default it works over xfs only
  • it requires package stratisd and stratis-cli
  • to get local repo for rhel8
baseurl=ftp://192.168.10.254/pub/rhel8/AppStream/
baseurl=ftp://192.168.10.254/pub/rhel8/BaseOS/
  • start its services by
  systemctl start stratisd.service
  systemctl enable stratisd.service
  • to create a Pool
stratis pool create pool-name Block-device-name1 block-device-name2
  • to check Pool
stratis pool list
  • to check block devices attached
stratis blockdev list
  • to create filesystem
stratis filesystem create name-of-pool-to-use filesystem-name1
  • to mount you can use Device name1
mount /stratis/pool-name/file-name /mnt/folder-name1
  • to extend pool storage / or attach another HDD
stratis pool add-data  pool-name disk-name
  • its fstab should be written as such that first stratis service should be up then only mount the device
device-path mount-folder-path format x-systemd.requires=stratisd.service  0 0

  • in format 0 0 defines backup and protection by company default structure

  • To create a snapshot(backup) of filesystem

stratis filesystem snapshot pool-name filesystem-name-to-make-snapshot-of file-name-of-snapshot
stratis filesystem snapshot pool1 file1 snap1
  • to find the created list of snapshots
stratis filesystem list
  • a snapshot is a copy of file system which can be allocated anywhere and it takes size from pool (maybe not sure)

Machine Learning

KNN CLASSIFICATION

  • KNN stands for K-Nearest Neighbour
  • K is a constant value
  • the coordinate to be classified calculates distance value to no of K different plotted data points nearest to itself
  • KNN works on Ecludien Distance Algorithm
  • the one with maximum points marked from nearby points is classified as answer
  • this method is too slow and fails for very huge amount of data

    Working

    • Find the distance from each point for the test point and sort them out to find the least distant points
  • the value of K should be odd and not 1 that is 3,5,7,9,...(RECOMMENDATION NOT COMPULSORY)
  • It can be used for the purpose of Regression as well

Regression

  • It is used when we have to find some specific value depending upon the features and values given
  • it uses linear graphical line plot to predict the value
  • it has many categories :
    1. Linear Regression ( y = mx + c)
    2. Polynomial Regression ( y = ax1 + a1x2**2 + a2x3**2)
    3. Logistic Regression

SVM & SVR

  • Support Vector Machine used in Classification
  • Support Vector Regression used in Regression
  • SVM creates a common support vector between two objects and analyze the new object based on location of new object

Notes

  • Google Brain Team - > open source,creator of tensor flow,deep learning team
  • Udacity - > company by google
  • slashml.blogspot.com - > website for Regression learning