Details
Description
Hi!
I've already created a thread into the citrix forums:
http://discussions.citrix.com/topic/367983-hp-smart-array-p411-xs65-sp1-hpsa-or-xenserver-broken-workaroundet-unsolved/
I used XenCenter, choose to create a new Virtual Disk Storage via Hardware HBA,
went on and after probing for luns, XenCenter shows me, what I was searching for: My just created LUN
It correctly asks me to format the drive... telling me very shortly, that it creates the storage and then nothig happens (I can see an error-message popping up, but it disappears after a fraction of a second - why?).
I got that from /var/log/messages:_
Aug 8 02:16:44 xenserver2 xapi: [ info|xenserver2|573 INET :::80|sm_exec D:6e2357871de1|xapi] Session.create trackid=73121ee6f98669714ed9fbedd1845604 pool=false uname= originator= is_local_superuser=true auth_user_sid= parent=trackid=9834f5af41c964e225f24279aefe4e49
Aug 8 02:16:45 xenserver2 kernel: [ 861.188406] scsi: host 0 channel 3 id 0 lun4194304 has a LUN larger than allowed by the host adapter
Aug 8 02:16:47 xenserver2 xapi: [ info|xenserver2|573 INET :::80|sm_exec D:6e2357871de1|xapi] Session.destroy trackid=73121ee6f98669714ed9fbedd1845604
Aug 8 02:17:29 xenserver2 xapi: [ info|xenserver2|589|Async.SR.probe R:3221c19bd6f1|dispatcher] spawning a new thread to handle the current task (trackid=e3dba2f5498accdf176f5b839daa7e7b)
Aug 8 02:17:29 xenserver2 xapi: [ info|xenserver2|589|sm_exec D:ec2c60b9195f|xapi] Session.create trackid=61c0a2ee1c954e5184793b2063b03e8b pool=false uname= originator= is_local_superuser=true auth_user_sid= parent=trackid=9834f5af41c964e225f24279aefe4e49
Aug 8 02:17:29 xenserver2 xapi: [ info|xenserver2|589|sm_exec D:ec2c60b9195f|xapi] Session.destroy trackid=61c0a2ee1c954e5184793b2063b03e8b
Aug 8 02:17:32 xenserver2 xapi: [ info|xenserver2|573 INET :::80|SR.create R:cbef48d054d6|storage_access] SR df8e256a-98cd-f81e-c93b-00d84f8eaab6 will be implemented by /services/SM/lvmohba in VM OpaqueRef:c78d1007-88bf-697c-2ec6-8fc1502ba401
Aug 8 02:17:32 xenserver2 xapi: [ info|xenserver2|573 INET :::80|sm_exec D:0938b0b67789|xapi] Session.create trackid=a2b308013133d26aff4b0a1cce5ad73f pool=false uname= originator= is_local_superuser=true auth_user_sid= parent=trackid=9834f5af41c964e225f24279aefe4e49
Aug 8 02:17:33 xenserver2 xapi: [ info|xenserver2|573 INET :::80|sm_exec D:0938b0b67789|xapi] Session.destroy trackid=a2b308013133d26aff4b0a1cce5ad73f
Aug 8 02:17:33 xenserver2 xapi: [error|xenserver2|573 INET :::80|SR.create R:cbef48d054d6|storage_access] Re-raising as SR_BACKEND_FAILURE_19 [ ; No such device [opterr=Device /dev/disk/by-id/scsi-3600508b1001c929c90b17a21452ed01b does not exist]; ]_
First of all:
Aug 8 02:16:45 xenserver2 kernel: [ 861.188406] scsi: host 0 channel 3 id 0 lun4194304 has a LUN larger than allowed by the host adapter
is totally nonsense of course. The LUN is smaller than 2 TB and has 24 disks - even after deleting the LUN and creating just a single one with 8x 72 GB disks produced that message.
I couldn't find anything about a limit-per-lun, just that the controller is limited, in worst case, to 100 drives (by 2 storage systems รก 25 drives on each of the 2 ports [2x2x25]).
Comeing back to the main problem:
Aug 8 02:17:33 xenserver2 xapi: [error|xenserver2|573 INET :::80|SR.create R:cbef48d054d6|storage_access] Re-raising as SR_BACKEND_FAILURE_19 [ ; No such device [*opterr=Device /dev/disk/by-id/scsi-3600508b1001c929c90b17a21452ed01b does not exist*]; ]
this was absolutely correct, because it is called cciss-36... and not scsi-36...:
Which now points me to 2 problems:
1. hpsa (3.4.6-170 - the newest one) or xenserver is giving out a wrong message about the LUN-size
2. whatever process is involved, is scrambling device-/lun-names.
I 'solved' my problem by just creating sr via console and not using /dev/disk/by-id... though using /dev/sda1 (created a partition table on sda):
[root@xenserver2 ~]# xe sr-create content-type=user device-config:device=/dev/sda1 shared=false type=lvm name-label=15ktest
30946b27-461b-811a-9dde-00ce50b504aa
Which worked without any problems. I'm currently benchmarking the stuff. It's not a blast, but dd gives me:
root@nagios:~# dd if=/dev/zero of=/dev/xvdb bs=1G count=100 oflag=direct
100+0 records in
100+0 records out
107374182400 bytes (107 GB) copied, 652.545 s, 165 MB/s
From within a debian 7 VM. I'm looking more forward to the IOPs, which I may test later on.
I will have the system for possible testing for some time, but may need a bit of time, because it's not installed anywhere and currently just in a testing/preparation state.
Regards
- Christof