Grokking DevOps and SRE Interviews
0% completed
Exercise: Reading the End of a Large File
A payment service writes 40 GB of logs a day into one file. During an incident, the on-call engineer runs tail -n 20 on that file and sees the last 20 lines immediately.
A script that reads the same file line by line from the start must read all 40 GB first. At 500 MB a second, that takes 80 seconds, because 40,000 MB divided by 500 MB a second is 80. A script that calls readlines() is worse, because it also tries to hold all 40 GB in memory.
The tail command avoids both costs on a regular file. It moves to the end and reads backward in fixed-size blocks until it has seen enough lines
.....
.....
.....
Like the course? Get enrolled and start learning!
Reading Progress
0%