1st phone interview
internal aws
search using solr
kafka
zookeeper
sort system?
data centers
scale
every phone is a client
- a deverloper said his binary is running slow. and need more hardware.
how do you know if we really need more hardware?
top
- if cpu usage is low, still seeing high load. how possible?
some processes in waiting state, such as i/o.
- if i see load 16 is it high?
not necessarily. if on 32 core machine it's fine.
main dev
devops sre
services automation
min support
mesos is used on compute side.
he is in stateful ervice.
lots of python code automate
on call once a quarter
produce a lot of code
accelerated
------
def is_palindrome(candidate):
candidate = candidate.lower().replace(' ', '')
candidate = [c for c in candidate if c.isalnum()]
return candidate == candidate[::-1]
for candidate in candidates:
print "{}: {}".format(candidate, is_palindrome(candidate)
>>> string = "Special $#! characters spaces 888323"
>>> ''.join(e for e in string if e.isalnum())
'Specialcharactersspaces888323'
=========================
2nd phone interview
# Given two non overlapping, sorted range pairs.
# A = [2,5], [8, 9], [11, 13]
# B = [1,7], [9, 11], [13,17]
# find the overlap between the two.
# [2,5],[9], [11]
# or
# 2,3,4,5,9
solution 1:
convert A, B into two sets with all integers.
do a intersection of A and B.
this will give 2nd answer.
solution 2:
def overlap([x,y], [a,b]:
if x > b or y < a:
return None
else:
return [max(x,a), min(y,b)]
def print_overlap(A, B):
res = []
for [x, y] in A:
for [a, b] in B:
if not overlap([x,y], [a,b]):
break
res.append(overlap([x,y], [a,b]))
------------------------
challenge projects
automation
dever doing devops currently
automation is way to go
solr cloud
what is the architecture of es?
how did you tune es?
lucene
open source api
filter out
proxy cloud
rate limiting
mesos 10k hosts
solr cloud is about 10k hosts, not all used for searching though.
- how did you implment access control?
TLS
distributed shared secrets
security is difficult to manage.
internal aws
search using solr
kafka
zookeeper
sort system?
data centers
scale
every phone is a client
- a deverloper said his binary is running slow. and need more hardware.
how do you know if we really need more hardware?
top
- if cpu usage is low, still seeing high load. how possible?
some processes in waiting state, such as i/o.
- if i see load 16 is it high?
not necessarily. if on 32 core machine it's fine.
main dev
devops sre
services automation
min support
mesos is used on compute side.
he is in stateful ervice.
lots of python code automate
on call once a quarter
produce a lot of code
accelerated
------
def is_palindrome(candidate):
candidate = candidate.lower().replace(' ', '')
candidate = [c for c in candidate if c.isalnum()]
return candidate == candidate[::-1]
for candidate in candidates:
print "{}: {}".format(candidate, is_palindrome(candidate)
>>> string = "Special $#! characters spaces 888323"
>>> ''.join(e for e in string if e.isalnum())
'Specialcharactersspaces888323'
=========================
2nd phone interview
# Given two non overlapping, sorted range pairs.
# A = [2,5], [8, 9], [11, 13]
# B = [1,7], [9, 11], [13,17]
# find the overlap between the two.
# [2,5],[9], [11]
# or
# 2,3,4,5,9
solution 1:
convert A, B into two sets with all integers.
do a intersection of A and B.
this will give 2nd answer.
solution 2:
def overlap([x,y], [a,b]:
if x > b or y < a:
return None
else:
return [max(x,a), min(y,b)]
def print_overlap(A, B):
res = []
for [x, y] in A:
for [a, b] in B:
if not overlap([x,y], [a,b]):
break
res.append(overlap([x,y], [a,b]))
------------------------
challenge projects
automation
dever doing devops currently
automation is way to go
solr cloud
what is the architecture of es?
how did you tune es?
lucene
open source api
filter out
proxy cloud
rate limiting
mesos 10k hosts
solr cloud is about 10k hosts, not all used for searching though.
- how did you implment access control?
TLS
distributed shared secrets
security is difficult to manage.
Comments
Post a Comment
https://gengwg.blogspot.com/