Score:0

changing queue parameter nr_request on NVMe dev

pe flag

We would like to use NVMes (SAMSUNG MZ1LB1T9HALS-00007) as a cache-tier in a Ceph System (17.2.0). The problem is that the stored data has a lot of small files and if hot-tiering is set to ~70% on the NVMes many slow-ops show up. Therefore, we thought of increasing ​​nr_requests​​ on the NVMes, i.e., the maximum number of read and write requests that can be queued at one time. However, the limit seems to be 1023:

echo "2046" > /sys/block/nvme0n1/queue/nr_requests
bash: echo: write error: Invalid argument

Used System: Kernel 5.15.0-60-generic, Ubuntu 22.04.1 LTS.

Is there a way to avoid this limitation or is the setup per se not a good idea?

I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.