Q: 3
A data scientist is training a deep learning model and notices slower than expected training times.
The data scientist alerts a system administrator to inspect the issue. The system administrator
suspects the disk IO is the issue.
What command should be used?
Options
Discussion
B . iostat is literally made for monitoring disk IO, which is exactly what the sysadmin suspects here. tcpdump would be for network, nvidia-smi for GPU, htop for processes and CPU. Pretty straightforward unless there's a trick I'm missing.
B . iostat is specific for checking disk IO bottlenecks. D (htop) might look helpful but it's more CPU/mem focused, not IO stats. Saw similar question in a practice test and the trap is picking D instead of B.
Nah, it's not D here-B is the tool for disk IO issues. htop tricks a lot of folks since it looks flashy, but it won't give detailed block device stats like iostat does. Pretty sure B is correct, anyone disagree?
B is the way to go here, since iostat directly shows disk IO bottlenecks. I remember seeing a similar scenario pop up in an exam simulation and it was always about matching the tool to the suspected hardware issue. The other options don't really give you block device stats. Pretty sure B is right, agree?
I actually thought D (htop) because it shows system resource usage in real-time, including IO wait. But now realizing it doesn’t break down disk IO specifics like iostat does. Maybe I’m missing something, but htop was my first guess. Disagree?
Its B for this, but if the model data was being loaded over NFS or some network mount, you'd probably need something like tcpdump or nload to really see network IO impact, not just disk. With normal local disks though, iostat is what you want. Anyone see an edge case where htop would matter more?
Call it B for this one. iostat gives disk IO stats directly, which is what you'd want since the admin suspects IO bottlenecks. tcpdump and nvidia-smi don't fit here. Pretty confident but open if anyone disagrees.
I remember a similar scenario from labs. in practice exams, pretty sure it's B here.
Option B since iostat directly shows disk IO performance, but if the workload is network-based or on a cloud volume, this might not catch everything. In some edge setups htop might hint at IO-wait spikes though.
B imo, iostat is made for monitoring disk IO directly. The others don't give block device stats so you'd miss the actual bottleneck. If it was a GPU thing then C might fit, but for disk issues it's gotta be B. Anyone see a reason to pick something else?
Be respectful. No spam.
