added additional ip-addresses to iscsi targets, how can we get RHEV-H to discover this?

Latest response

I just added an additional ip-address to each of my iscsi targets, and now would very much like my RHEV-H hypervisors to discover the new paths and start using them. Any ideas for how to achieve that?

I tried going to an SD, edit it, discover targets, input new ip-address, chap username = initiatorname of the "use host" value, and a password. This works for the single host that is the "use host"/SPM, but the other hypervisors doesn't pick up this. I've tried setting them in maintenance/activating, and also rebooting them.. But no new paths are discovered for any other host than the SPM.

Any ideas?

 

How I want it to look:

# multipath -ll

36005076802808538a00000000000009d dm-10 IBM,2145
size=2.0T features='0' hwhandler='0' wp=rw
|-+- policy='round-robin 0' prio=50 status=active
| |- 1:0:0:4 sdf  8:80   active ready running
| `- 4:0:0:4 sdz  65:144 active ready running
`-+- policy='round-robin 0' prio=10 status=enabled
  |- 3:0:0:4 sdr  65:16  active ready running
  `- 5:0:0:4 sdah 66:16  active ready running
<snip>

# iscsiadm -m session
tcp: [1] 192.168.138.4:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node1
tcp: [2] 192.168.129.4:3260,2003 iqn.1986-03.com.ibm:sn.151745642
tcp: [3] 192.168.138.5:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node2
tcp: [4] 192.168.138.6:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node1
tcp: [5] 192.168.138.7:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node2
 

How it looks on non-SPM hosts:

# multipath -ll

36005076802808538a00000000000009d dm-10 IBM,2145
size=2.0T features='0' hwhandler='0' wp=rw
|-+- policy='round-robin 0' prio=50 status=active
| `- 1:0:0:4 sdf 8:80  active ready running
`-+- policy='round-robin 0' prio=10 status=enabled
  `- 3:0:0:4 sdr 65:16 active ready running
<snip>

# iscsiadm -m session
tcp: [1] 192.168.138.4:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node1
tcp: [2] 192.168.129.4:3260,2003 iqn.1986-03.com.ibm:sn.151745642
tcp: [3] 192.168.138.5:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node2
 

ja
Log in to join the conversation

Responses

rhn-support-anande's picture Red Hat Guru 1434 points
19 June 2013 4:28 PM

I see that you have 4 paths:

1:0:0:4 sdf  8:80   active ready running

4:0:0:4 sdz  65:144 active ready running

3:0:0:4 sdr  65:16  active ready running

5:0:0:4 sdah 66:16  active ready running

If I am getting this correctly, for every path you have one lun:
  • tcp: [1] 192.168.138.4:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node1
  • tcp: [3] 192.168.138.5:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node2
  • tcp: [4] 192.168.138.6:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node1
  • tcp: [5] 192.168.138.7:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node2
 
and one additional which is part of a different IBM storage:
  • tcp: [2] 192.168.129.4:3260,2003 iqn.1986-03.com.ibm:sn.151745642
 
On the non-spm host, I see that only 2 paths and 3 targets are being discovered
  • 1:0:0:4 sdf 8:80  active ready running
  • 3:0:0:4 sdr 65:16 active ready running
  • tcp: [1] 192.168.138.4:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node1
  • tcp: [2] 192.168.129.4:3260,2003 iqn.1986-03.com.ibm:sn.151745642
  • tcp: [3] 192.168.138.5:3260,1 iqn.1986-03.com.ibm:2145.mro-adc-v7000u.node2
  •  Can you confirm that from a non-SPM host, you can login to the missing targets above ?
  •  Also, from confirm that the mapping of the non-SPM hosts has been done properly for these targets.
  • If the issue still persists, I would recommend you to open a support ticket with us,so that we can dig into this deeper. We might also need to analyze the TCPDUMP for checking where exactly is the issue.
ja Guru 1340 points
19 June 2013 7:29 PM

 > Can you confirm that from a non-SPM host, you can login to the missing targets above ?

All my hosts has initiatorname=iqn.1994-05.com.redhat:$hostname.$domainname and they all have the same chap secret. They should all be able to log in to the targets using this initiatorname+chap secret on all paths, but I don't know how I can confirm this. I can't discover these new addresses, but that's probably because the /etc/iscsi/iscsid.conf doesn't contain the chap secret and I don't know how RHEV-H manages this..

# iscsiadm -m discovery -t st -p 192.168.138.6
iscsiadm: Connection to Discovery Address 192.168.138.6 closed
iscsiadm: Login I/O error, failed to receive a PDU
<repeats>

iscsiadm: Could not perform SendTargets discovery: encountered iSCSI login failure
 

> Also, from confirm that the mapping of the non-SPM hosts has been done properly for these targets.

The target has the exact same config for all LUNs and initiators. And all hosts can already log in to both of the targets on the first ip address. I have no target side configuration that treats each of the ip addresses for the same target differently.

> If the issue still persists, I would recommend you to open a support ticket with us,so that we can dig into this deeper. We might also need to analyze the TCPDUMP for checking where exactly is the issue.

I have opened a support ticket (00893683), but it would be great to resolve this in the open forum. I'm pretty sure this isn't a rhev-bug, but rather a missing documentation or knowledge on my part.

Close

Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.