No. 0451 min read
Splunk versus ELK. Put the holy war down, folks.
I used Splunk at Raymond James and ELK at Sprint and Jabil.
#operations
9 stories about operations. Newest on top.
No. 0391 min read
Logs are for the poor bastard awake at 3 a.m.
Tell me what failed, what might already have happened, and what I can safely do next.
No. 0382 min read
Cron started the command. What the hell else did you expect?
Cron starts commands when the clock matches, folks.
No. 0371 min read
Atomic rename. Read the damned small print.
Unix rename can publish a finished file or claim a queue item.
No. 0341 min read
The error queue is unfinished accounting. Go get it.
Your parser rejected the record.
No. 0151 min read
No customer impact. You think nothing happened?
Sprint's D3884 finished early and under budget with no customer service impact.
No. 0132 min read
Runbooks for a smaller team. Write down the damned part you know.
Sprint's mediation ops team went from twelve people to four without losing coverage.
No. 0091 min read
Fourteen years at five nines. Yes, we patched them.
Fifty servers, fourteen years, 99.999% availability.
No. 0072 min read
NT to Linux at sixteen. Yes, I still had homework.
At sixteen I moved Advertisnet's services from Windows NT to Linux/BSDi.