Differences

This shows you the differences between two versions of the page.

--- nndocs:srp [2025/03/14 05:31] – naptastic
+++ nndocs:srp [2025/12/25 00:51] (current) – [Dependencies] Note: using port 2 doesn't seem to work naptastic
@@ Line 14: / Line 14: @@
   fe800000000000005849560e59150301
-lastly change fe80 to 0000 and create the ACL:
+Initiator ACLs start with all 0's. Targets start with fe80.
   /srpt> ib.fe800000000000005849560e53b70b01/acls create ib.00000000000000005849560e59150301
   Created Node ACL for ib.00000000000000005849560e59150301
   Created mapped LUN 0.
+A Linux SRP target is always visible from all InfiniBand partitions.
 ====Dependencies====
   apt install srptools
+Do **NOT** set srp_daemon loose without using the -o flag! It will flood dmesg on both the initiator and the target!
 Find targets to connect to:
@@ Line 32: / Line 36: @@
   * -c means "emit target information in a format we can use later"
   * -p 1 means "only scan on HCA port 1" so obviously change this if you are initiating from port 2...
+Note: in my experimentation, port 2 is not actually usable. A target created on a dual-port card always answers on port 1 even if the associated GUID is for port 2, and an initiator on port 2 always actually tries to use port 1. The answer is probably "use SR-IOV".
+====Configuration====
+It is **critical** that you edit /etc/srp_daemon.conf as soon as you have a list of targets and disallow connections to anything except the targets you want. The default file is well commented.
+To connect to a target listed by srp_daemon, write it to the appropriate add_target file in /sys/class/infiniband_srp. Here's how shark gets its swap ramdisk from southpark:
+  [root]@[shark][~]# echo 'id_ext=5849560e53b70b01,ioc_guid=5849560e53b70b01,dgid=fe800000000000005849560e53b70b01,pkey=ffff,service_id=5849560e53b70b01' > \
+  /sys/class/infiniband_srp/srp-ibp14s0f0-1/add_target
+  [root]@[shark][~]# dmesg
+    (...snip...)
+  [2719206.378801] scsi host8: SRP.T10:5849560E53B70B01
+  [2719206.379439] scsi 8:0:0:0: Direct-Access     LIO-ORG  swap             4.0  PQ: 0 ANSI: 6
+  [2719206.380206] sd 8:0:0:0: Attached scsi generic sg5 type 0
+  [2719206.380337] sd 8:0:0:0: [sdd] 33554432 512-byte logical blocks: (17.2 GB/16.0 GiB)
+  [2719206.380376] scsi host8: ib_srp: new target: id_ext 5849560e53b70b01 ioc_guid 5849560e53b70b01 pkey ffff service_id 5849560e53b70b01 sgid fe80:0000:0000:0000:5849:560e:5915:0301 dgid fe80:0000:0000:0000:5849:560e:53b7:0b01
+  [2719206.380380] sd 8:0:0:0: [sdd] Write Protect is off
+  [2719206.380384] sd 8:0:0:0: [sdd] Mode Sense: 43 00 00 08
+  [2719206.380452] sd 8:0:0:0: [sdd] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
+  [2719206.395027] sd 8:0:0:0: [sdd] Preferred minimum I/O size 512 bytes
+  [2719206.395030] sd 8:0:0:0: [sdd] Optimal transfer size 4294967288 logical blocks > dev_max (65535 logical blocks)
+  [2719206.425809] sd 8:0:0:0: [sdd] Attached SCSI disk
 Lazy benchmarking seems good:
-  # dd if=/dev/sdb of=/dev/null bs=4M
+  [root]@[shark][~]# dd if=/dev/sdb of=/dev/null bs=4M
 +0 records in
 +0 records out
   17179869184 bytes (17 GB, 16 GiB) copied, 5.38771 s, 3.2 GB/s
-  [root]@[shark][00:23:02][~]# dd if=/dev/zero of=/dev/sdb bs=4M
+  [root]@[shark][~]# dd if=/dev/zero of=/dev/sdb bs=4M
   dd: error writing '/dev/sdb': No space left on device
 +0 records in
@@ Line 46: / Line 73: @@
   17179869184 bytes (17 GB, 16 GiB) copied, 13.7431 s, 1.3 GB/s
-...so I'm using some for swap. We'll see how long it takes me to induce a kernel panic. ;)
+====Logout====
+"Delete the port" sounds pretty destructive, but this actually is the graceful way to close the connection.
+  # echo 1 > /sys/class/srp_remote_ports/ [tab tab tab] /delete