Nova-compute connects to all of the iscsi target

[Hiroyuki Eguchi]

Nova-compute tries to connect to all of the iscsi target which discovered when using multhpath.
As a result , a lot of unnecessary iscsi sessions occur on the nova-compute.

We have to choose correct targets from output of iscsiadm discovery by iscsi_properties of the volume we try to connect.

Steps of attaching multipath device is like this:

1. discover iscsi targets
2. get all active iscsi sessions
3. compare 1. with 2. and connect to all iscsi targets not connected

It should be like this:

1. discover iscsi targets
2. choose correct targets from output of discovery
3. get all active iscsi sessions
4. compare 2. with 3. and connect to the iscsi target if it is not connected

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/kilo
Review: https://review.openstack.org/298084

[Sylvain Bauza (sylvain-bauza)]

Could you please reproduce the issue against master to see if the bug is still valid ?

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/299126

[Hiroyuki Eguchi (h-eguchi)]

This bug is still valid in master.
However, this source file have been migrated as os-brick since liberty.
I've commited a patch against master of os-brick.

[Hiroyuki Eguchi (h-eguchi)]

There are two types of iSCSI multipath devices.

type1) shares the same iqn between multiple portals.(e.g. NEC Storage I'm using)

# iscsiadm -m session
10.XXX.XXX.XX6:3260,6 iqn.2001-03.jp.nec:storage01:ist-m000-sn-0000000j1bn00458.lx-openstack1-0088.target0070
10.XXX.XXX.XX1:3260,1 iqn.2001-03.jp.nec:storage01:ist-m000-sn-0000000j1bn00458.lx-openstack1-0088.target0070

type2) use different iqns on different portals.(e.g. EMC Storage)

# sudo iscsiadm -m session
10.XXX.XXX.XX4:3260,4 iqn.1992-04.com.emc:cx.fcn00133400150.a7
10.XXX.XXX.XX8:3260,8 iqn.1992-04.com.emc:cx.fcn00133400150.b7

This bug is only valid against type1.
In case of type1, client can found out which targets is correct by iqn stored in iscsi_properties,
so it is unnecessary to connect to all of the targets which discovered.

So we have to take into account this difference.
I'm going to modify a patch later.

[OpenStack Infra (hudson-openstack)]

Change abandoned by Dave Walker (<email address hidden>) on branch: stable/kilo
Review: https://review.openstack.org/298084
Reason:
stable/kilo closed for 2015.1.4

This release is now pending its final release and no freeze exception has
been seen for this changeset. Therefore, I am now abandoning this change.

If this is not correct, please urgently raise a thread on openstack-dev.

More details at: https://wiki.openstack.org/wiki/StableBranch

[OpenStack Infra (hudson-openstack)]

Change abandoned by Sean McGinnis (<email address hidden>) on branch: master
Review: https://review.openstack.org/299126
Reason: This review is > 4 weeks without comment, and failed Jenkins the last time it was checked. We are abandoning this for now. Feel free to reactivate the review by pressing the restore button and leaving a 'recheck' comment to get fresh test results.

[Sean McGinnis (sean-mcginnis)]

Unassigning due to no activity for > 6 months.