Created attachment 958022 [details] logs Description of problem: see bz #1159637 The operation of detach-attach an iSCSi storage domain does not login to iSCSi sessions as it should (on my hosts 4 iqn's are configured) Upon attach the receiving host is logged to 3 session out of 4 iqn's ,furthermore, it is somehow logged in to each session twice(one nic only): [root@dhcp-2-51 ~]# lspci | grep Ether 00:19.0 Ethernet controller: Intel Corporation 82567LM-3 Gigabit Network Connection (rev 02) [root@dhcp-2-51 ~]# iscsiadm -m session tcp: [17] 10.35.160.105:3260,2 iqn.1992-04.com.emc:cx.ckm00121000438.a7 tcp: [18] 10.35.160.105:3260,2 iqn.1992-04.com.emc:cx.ckm00121000438.a7 tcp: [19] 10.35.160.106:3260,3 iqn.1992-04.com.emc:cx.ckm00121000438.b6 tcp: [20] 10.35.160.106:3260,3 iqn.1992-04.com.emc:cx.ckm00121000438.b6 tcp: [21] 10.35.160.107:3260,4 iqn.1992-04.com.emc:cx.ckm00121000438.b7 tcp: [22] 10.35.160.107:3260,4 iqn.1992-04.com.emc:cx.ckm00121000438.b7 tcp: [23] 10.35.160.104:3260,1 iqn.1992-04.com.emc:cx.ckm00121000438.a6 see also sessions.log file(it is the output of getDeviceList) Version-Release number of selected component (if applicable): vt10 How reproducible: 100% Steps to Reproduce: * Have a storage group with 4 iqn's 1.(on first dc) Create iSCSi domain 2.(on first dc) Create NFS Domain 3.(on second dc) Craete NFS domain 4. detach iSCSi domain and attach it to second dc Actual results: Host receiving the domain has connections duplicity Expected results: no duplicity, the host is logged in to all sessions needed to see the iSCSi domain Additional info:
Liron/Greg, is this related to bug 1176402?
(In reply to Allon Mureinik from comment #1) > Liron/Greg, is this related to bug 1176402? Unsure. I'm figuring out (with Ori) if/how I can reproduce the issue and won't know more until then.
(In reply to Greg Padgett from comment #2) > (In reply to Allon Mureinik from comment #1) > > Liron/Greg, is this related to bug 1176402? > > Unsure. I'm figuring out (with Ori) if/how I can reproduce the issue and > won't know more until then. Can you please update of the status of this issue?
(In reply to Yaniv Dary from comment #3) > (In reply to Greg Padgett from comment #2) > > (In reply to Allon Mureinik from comment #1) > > > Liron/Greg, is this related to bug 1176402? > > > > Unsure. I'm figuring out (with Ori) if/how I can reproduce the issue and > > won't know more until then. > > Can you please update of the status of this issue? Hi Yaniv, I can't reproduce it in my dev environment, so I'm getting a virt host that can connect to the tested storage and will debug/fix it from there. (Same applies for bug 1164578 as well)
3.5.1 is already full with bugs (over 80), and since none of these bugs were added as urgent for 3.5.1 release in the tracker bug, moving to 3.5.2
Created attachment 1035950 [details] logs reproduced on rhevm 3.5 vt15.4. [root@fury66 ~]# iscsiadm -m session tcp: [1] 10.35.160.107:3260,4 iqn.1992-04.com.emc:cx.ckm00121000438.b7 (non-flash) tcp: [2] 10.35.160.107:3260,4 iqn.1992-04.com.emc:cx.ckm00121000438.b7 (non-flash) tcp: [3] 10.35.160.105:3260,2 iqn.1992-04.com.emc:cx.ckm00121000438.a7 (non-flash) tcp: [4] 10.35.160.106:3260,3 iqn.1992-04.com.emc:cx.ckm00121000438.b6 (non-flash) tcp: [4] 10.35.160.106:3260,3 iqn.1992-04.com.emc:cx.ckm00121000438.b6 (non-flash) tcp: [5] 10.35.160.104:3260,1 iqn.1992-04.com.emc:cx.ckm00121000438.a6 (non-flash) please be aware-> two of the hosts have one Ethernet controller each, this does not reproduce on machines with two nics or more
This bug is Verfied on latest 3.6 rhevm build (12)
Verified on latest oVirt 3.5 and latest Rhevm 3.6. note: iscsi connection are not doubled any more but the final result of this flow differs from 3.5 to 3.6. while in 3.6 'hypervisor A' which is detaching iscsi domain and "left" with only file storage has no iscsi connection upon detach. in 3.5 'hypervisor A' left with redundant connections upon detach: root@el71-host01 ~ $ iscsiadm -m session tcp: [11] 10.35.160.106:3260,3 iqn.1992-04.com.emc:cx.ckm00121000438.b6 (non-flash) tcp: [6] 10.35.160.107:3260,4 iqn.1992-04.com.emc:cx.ckm00121000438.b7 (non-flash) tcp: [8] 10.35.160.105:3260,2 iqn.1992-04.com.emc:cx.ckm00121000438.a7 (non-flash)
We didn't get 3.5.5 rhevm build yet - therefore it wasn't verified on the relevant build. Back to on_qa to verify once we will have 3.5.5 downstream build.
Verified on 3.6 based on comment #8
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2016-0362.html