NVME-ATTACH-NS(1)
π nvme-attach-ns β like, the NVMe manual, but make it make sense π
okay so bestie, you're looking at nvme-attach-ns from the NVMe Manual β and heads up, this one's a legacy queen π. She's deprecated, but she still shows up to the party for the old scripts. Let's get into it β¨
β¨ NAME β¨
nvme-attach-ns β send the NVMe attach namespace command and get the result back. (deprecated, bestie, I'm so serious π)
π¦ SYNOPSIS π¦
nvme [<global-options>] attach-ns <device>
[--namespace-id=<nsid> | -n <nsid>]
[--controllers=<ctrl-list,> | -c <ctrl-list,>]
That's the whole vibe check: you give it a <device>, then pick your namespace with -n / --namespace-id=<nsid> and your controllers with -c / --controllers=<ctrl-list,>. Slay π«Ά
πΈ DESCRIPTION πΈ
So here's the tea β: nvme-attach-ns is a deprecated alias for nvme-ns-attach(1). Like, it's the same girl, different name tag. It's been kept around specifically for scripts that were written against nvme-cli versions older than 3.0 β we're not gonna break your old code, that's not the energy here π
The options and the behavior are identical. Literally the only thing that got kept is the top-level command name. That's it. That's the whole difference π
New scripts, though? Bestie, please use nvme ns attach instead. This alias prints a deprecation warning on stderr β which, yeah, it's gonna be loud about it π β and it may be removed in a future release. So don't get too attached (pun intended π
).
For the full option reference, go read nvme-ns-attach(1). She's got the whole menu β¨
π GLOBAL OPTIONS π
These ones live up at the top-level nvme command, but they're totally available to this subcommand too. Group-chat access, iykyk π«Ά
--dry-runβ prints out the command that would be executed, but doesn't actually execute it. A little dress rehearsal, no commitment π--no-ioctl-probingβ turns off probing for 64-bit IOCTL support. Just... no probing today, thanks π--no-retriesβ disables the retry logic on transient errors. One shot, no second chances π-o <fmt>,--output-format=<fmt>β sets the reporting format tonormal,tabular,json, orbinary. Only one output format at a time, bestie β we can't have you being greedy π--output-format-version=<version>β picks the output format version. Version1uses the original field naming, while version2(the default) gives you more consistent and script-friendly field names. Any other version gets rejected, periodt π«--quietβ suppresses informational messages. Errors still get reported (obviously, we're not hiding the important stuff), and--quietcancels any--verbosefor the library log level. It's giving "I said what I said" π--set-options=<key=value>[,<key=value>...]β sets libnvme library options for this invocation. The supported keys areforce-4k,mi-probe-enabled,test-base-dir, andtest-sysfs-dir. Mix and match, comma-separated, go off π--timeout=<ms>β sets the timeout for the command in milliseconds.infiniteor-1waits with no time limit at all (0xffffffff). Patience is a virtue, I guess π-v,--verboseβ bumps up the level of detail in the output. You can specify it multiple times to get even more verbose. The way she just keeps going ππ₯
π‘ Little note, bestie: these options can also be set as machine-wide defaults in nvme-cli.conf(5). But a command-line flag always overrides the file β command line is main character energy, fr fr π
π EXAMPLES π
No Examples
Yeah. Nothing. Not a single one. Iconic, honestly πβ¨
π§ NVME π§
Part of the nvme-user suite π
β SEE ALSO β
nvme-ns-attach(1)β the un-deprecated one, the one you should actually be using, the main character π«Ά
β οΈ ok final reminder bestie: this whole page is deprecated. It still works, it still behaves exactly like nvme-ns-attach(1), but it's gonna warn you on stderr every time and it might vanish in a future release. Migrate your scripts and thank me later π
β¨
that's the page, babes β brought to you by the NVMe manual, dated 10/30/2026, and no, I did not forget the date, I would never ππ
π¦ where this page comes from
- arch rolling: nvme-cli 3.2-1 Β· GPL-2.0-or-later
- π girlified by deepseek-ai/DeepSeek-V4.1-Flash-1017 on : a language model's rewrite of the original, under its terms