All Questions
19
questions
0
votes
1
answer
44
views
Is my SSD dead? can't mount. I/O errors, 29k+ fsck errors
So I work with a couple of people and we ssh into a Linux server for work. There are several SSDs, and one of the non-boot Samsung 870 SSDs crashed the other day in the morning. This led to an ...
0
votes
1
answer
286
views
FIO reports slower sequential read than the advertised NVMe SSD read bandwidth
TLDL
For the very simple sequential read, FIO reports is much slower than the NVMe SSD sequentail read capability.
Main Text
Hello everyone,
I have been facing an issue while trying to achieve the ...
0
votes
1
answer
2k
views
Isolating I/O issue with NVME or hardware?
Hardware:
Samsung 980 PRO M.2 NVMe SSD (MZ-V8P2T0BW) (2TB)
Beelink GTR6, with the SSD in the NVMe slot
Since the hardware arrived, I've installed Ubuntu Server on it as well as a bunch of services (...
1
vote
1
answer
1k
views
What is the impact of different Linux I/O APIs on performance benchmarking of NVMe SSDs? (libaio vs SPDK vs io_uring)
I am benchmarking NVMe SSDs on my Linux server, with the aim to achieve the IOPS, BW and Latency values as mentioned in the product specifications.
I'm using FIO as workload generator, and used libaio ...
2
votes
1
answer
423
views
SSD performance different results in dd and gnome-disk
I wanted to measure my SSD read/write performance and found some recommendations for using dd and gnome-disks. (from my understanding, hdpart is not relevant since it's not using the disk itself but ...
1
vote
0
answers
213
views
NVME slower on host than in VM
I first noticed some strange performance hits when I switched from Vagrant to Docker for my local development setup.
After investigating why my databases were slower in this new environment, I came to ...
0
votes
0
answers
64
views
Modify the generic block layer in a Linux system
For an university project, I need to modify a kernel layer in a Linux operating system. After modification I should prove performance improvements.
Concrete I should modify the generic block layer and ...
1
vote
0
answers
1k
views
Disk IO bottleneck on Linux server
I've been facing Disk IO bottleneck on one of my streaming server, but couldn't identify the limiting factor. The issue might be on RAID card or in SSD due to which streaming application goes for the ...
2
votes
2
answers
3k
views
NVMe disk shows 80% io utilization, partitions show 0% io utilization
I have a CentOS 7 server (kernel 3.10.0-957.12.1.el7.x86_64) with 2 NVMe disks with the following setup:
# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
nvme0n1 259:0 0 477G 0 ...
2
votes
1
answer
4k
views
FIO reporting slow write speeds while DD reports fast ones
According to most disk benchmarks, sequential write speeds are typically very close to the read speeds. Mine are in the ~500MB/s range.
$ dd if=/dev/zero of=tempfile bs=1M count=5120 conv=notrunc ...
1
vote
1
answer
829
views
How to deal with blocking I/O reducing the performance to a crawl?
I have a computer which is experiencing some serious I/O issues.
Software: Kali-rolling, Linux x86-desktop-1 4.12.0-kali2-amd64 #1 SMP Debian 4.12.13-1kali2 (2017-10-03) x86_64 GNU/Linux. But I've ...
3
votes
0
answers
903
views
Interpreting iozone results
This is using a Samsung 960 Pro on Arch linux x64, linux 4.9.8-1, iozone3_465 (latest tar)
For example, the last line:
kB reclen write rewrite read reread read write read ...
24
votes
4
answers
20k
views
Monitor temperature/IO of NVMe drive
Is there a way to monitor temperature or reads/writes of and NVMe drive (in this case Intel 750). hdparm, udisksctl, smartctl, and hddtemp all seem to lack this capability, google searches have been ...
2
votes
1
answer
3k
views
hdparm: SECURITY_SET_PASS: Input/output error
I'm trying to use hdparm to secure erase an SSD as suggested here.
Following the directions in that guide has worked okay so far, but I can't seem to resolve this error:
~ $ sudo hdparm --user-...
0
votes
0
answers
91
views
SSD as write target during reads
I have a real-time application which always writes data to a RAID 10 HDD configuration in a sequential manner at around ~50 MB/s. It uses mmap and writing should never be blocked due to waiting I/O.
...