1NVME-CONNECT(1) NVMe Manual NVME-CONNECT(1)
2
6 nvme-connect - Connect to a Fabrics controller.
7
9 nvme connect
10 [--transport=<trtype> | -t <trtype>]
11 [--nqn=<subnqn> | -n <subnqn>]
12 [--traddr=<traddr> | -a <traddr>]
13 [--trsvcid=<trsvcid> | -s <trsvcid>]
14 [--host-traddr=<traddr> | -w <traddr>]
15 [--hostnqn=<hostnqn> | -q <hostnqn>]
16 [--nr-io-queues=<#> | -i <#>]
17 [--nr-write-queues=<#> | -W <#>]
18 [--nr-poll-queues=<#> | -P <#>]
19 [--queue-size=<#> | -Q <#>]
20 [--keep-alive-tmo=<#> | -k <#>]
21 [--reconnect-delay=<#> | -c <#>]
22 [--ctrl-loss-tmo=<#> | -l <#>]
23
25 Create a transport connection to a remote system (specified by --traddr
26 and --trsvcid) and create a NVMe over Fabrics controller for the NVMe
27 subsystem specified by the --nqn option.
28
30 -t <trtype>, --transport=<trtype>
31 This field specifies the network fabric being used for a
32 NVMe-over-Fabrics network. Current string values include:
33 ┌──────┬────────────────────────────┐
35 │Value │ Definition │
36 ├──────┼────────────────────────────┤
37 │rdma │ The network fabric is an │
38 │ │ rdma network (RoCE, iWARP, │
39 │ │ Infiniband, basic rdma, │
40 │ │ etc) │
41 ├──────┼────────────────────────────┤
42 │fc │ WIP The network fabric is │
43 │ │ a Fibre Channel network. │
44 ├──────┼────────────────────────────┤
45 │loop │ Connect to a NVMe over │
46 │ │ Fabrics target on the │
47 │ │ local host │
48 └──────┴────────────────────────────┘
49 -n <subnqn>, --nqn <subnqn>
51 This field specifies the name for the NVMe subsystem to connect to.
52 -a <traddr>, --traddr=<traddr>
54 This field specifies the network address of the Controller. For
55 transports using IP addressing (e.g. rdma) this should be an
56 IP-based address (ex. IPv4).
57 -s <trsvcid>, --trsvcid=<trsvcid>
59 This field specifies the transport service id. For transports using
60 IP addressing (e.g. rdma) this field is the port number. By
61 default, the IP port number for the RDMA transport is 4420.
62 -w <traddr>, --host-traddr=<traddr>
64 This field specifies the network address used on the host to
65 connect to the Controller.
66 -q <hostnqn>, --hostnqn=<hostnqn>
68 Overrides the default Host NQN that identifies the NVMe Host. If
69 this option is not specified, the default is read from
70 /etc/nvme/hostnqn first. If that does not exist, the autogenerated
71 NQN value from the NVMe Host kernel module is used next. The Host
72 NQN uniquely identifies the NVMe Host.
73 -i <#>, --nr-io-queues=<#>
75 Overrides the default number of I/O queues create by the driver.
76 -W <#>, --nr-write-queues=<#>
78 Adds additional queues that will be used for write I/O.
79 -P <#>, --nr-poll-queues=<#>
81 Adds additional queues that will be used for polling latency
82 sensitive I/O.
83 -Q <#>, --queue-size=<#>
85 Overrides the default number of elements in the I/O queues created
86 by the driver.
87 -k <#>, --keep-alive-tmo=<#>
89 Overrides the default keep alive timeout (in seconds).
90 -c <#>, --reconnect-delay=<#>
92 Overrides the default delay (in seconds) before reconnect is
93 attempted after a connect loss.
94 -l <#>, --ctrl-loss-tmo=<#>
96 Overrides the default controller loss timeout period (in seconds).
97
99 · Connect to a subsystem named
100 nqn.2014-08.com.example:nvme:nvm-subsystem-sn-d78432 on the IP4
101 address 192.168.1.3. Port 4420 is used by default:
102 # nvme connect --transport=rdma --traddr=192.168.1.3 \
104 --nqn=nqn.2014-08.com.example:nvme:nvm-subsystem-sn-d78432
105 SEE ALSO
107 nvme-discover(1) nvme-connect-all(1)
109
111 This was co-written by Jay Freyensee[1] and Christoph Hellwig[2] for
112 Keith Busch[3].
113
115 Part of the nvme-user suite
116
118 1. Jay Freyensee
119 mailto:james.p.freyensee@intel.com
120 2. Christoph Hellwig
122 mailto:hch@lst.de
123 3. Keith Busch
125 mailto:keith.busch@intel.com
126NVMe 01/08/2019 NVME-CONNECT(1)