Re: Logical Unit Not Ready, Initializing Cmd. Required
On May 16, 10:41 pm, Konrad Rzeszutek [EMAIL PROTECTED] wrote: So the block disks show up fine. That means iSCSI works just fine. starting multipath up solved the problem You only see one disk? Either way, are you running multipath on your machine? If you are, what is the multipath -ll output? Anyway, I am sending the multipath output in the configuration that works # multipath -ll 3600508b300920350a640cb720bce000edm-0 COMPAQ,MSA1000 VOLUME [size=838G][features=1 queue_if_no_path][hwhandler=1 hp_sw] \_ round-robin 0 [prio=4][active] \_ 13:0:0:1 sdb 8:16 [active][ready] Sometimes, depending mainly on the path_checker setting, the device fails. Once I had a segfault in multipathd. May 17 19:41:40 hostname multipathd: sdb: tur checker reports path is down May 17 19:41:51 hostname multipathd: sdb: tur checker reports path is down May 17 19:42:01 hostname multipathd: sdb: tur checker reports path is down May 17 19:42:07 hostname multipathd: 3600508b3009203503b2fc2c200040011: stop event checker thread May 17 19:42:07 hostname kernel: multipathd[13092]: segfault at 0012 rip 2b0b549d541d rsp 7fff5682d1b0 error 4 May 17 19:42:07 hostname multipathd: error calling out /sbin/scsi_id -g -u -s /block /sda May 17 19:42:07 hostname multipathd: sdb: checker msg is readsector0 checker reports path is down May 17 19:42:07 hostname multipathd: 3600508b3009203503b2fc2c200040011: event checker started In the present configuration the device sometimes goes away but returns quickly May 18 17:52:14 hostname kernel: sd 13:0:0:1: Device not ready: 6: Current: sense key: Not Ready May 18 17:52:14 hostname kernel: Additional sense: Logical unit not ready, cause not reportable May 18 17:52:14 hostname kernel: end_request: I/O error, dev sdb, sector 4775 May 18 17:52:14 hostname kernel: device-mapper: dm-multipath: Failing path 8:16. May 18 17:52:14 hostname multipathd: 8:16: mark as failed May 18 17:52:14 hostname multipathd: 3600508b300920350a640cb720bce000e: remaining active paths: 0 May 18 17:52:19 hostname multipathd: sdb: hp_sw checker reports path is ghost May 18 17:52:19 hostname kernel: device-mapper: hp-sw: Flushing I/O May 18 17:52:19 hostname kernel: device-mapper: hp-sw: sending START_STOP_UNIT command May 18 17:52:19 hostname multipathd: 8:16: reinstated May 18 17:52:19 hostname multipathd: 3600508b300920350a640cb720bce000e: remaining active paths: 1 The next step for enterprise storages, such as the HP StorageWorks is to issue a START SCSI command - which is what you should see if you are using multipath. The path checker would start the LUN. Yes that was the original problem. There was no multipath running and HP StorageWorks array would not start. When multipath was started the problem went away. Thank you for the advice. I have multipath 0.4.7 running the default configuration (no /etc/ multipath.conf) --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-iscsi@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/open-iscsi -~--~~~~--~~--~--~---
Re: Logical Unit Not Ready, Initializing Cmd. Required
On Thu, May 15, 2008 at 10:07:32PM -0700, sinysee wrote: Hello, I am trying to use open-iscsi to connect to an FC storage via an ATTO IPBridge 2700C. This is not a planned setup, but an emergency attempt to regain access to the data after an FC switch failure. I am using SLES 10.1 on the host. When trying to connect to the target I get through the discovery stage successfully. When I log in, iscsiadm returns success, but there are messages like sd 17:0:0:1: Device not ready: 6: Current: sense key: Not Ready Additional sense: Logical unit not ready, cause not reportable in the kernel log and the device never gets available. What do you mean by never gets available? Can you attach the full dmesg? Is it that the block device (/dev/sdX) that is unavailable or the multipath device (/dev/dm-XX)? Is it all of them or just one or two? I got the tcpdump from the host and it shows that the host repeatedly sends Test Unit Ready requests and receives Logical Unit Not Ready, Initializing Cmd. Required (0x0402) in response. Mode Sense and Read Capacity get meaningful responses. Then the host does a Read request and gets Logical Unit Not Ready, Cause Not Reportable (0x0400). Is that for each of the block disks that showed up when you logged in or just for some of them? If it is the latter than is expected with Active-Passive targets which you might have. .. snip.. Having very little prior experience with iSCSI I am not sure what Initializing Cmd. Required is supposed to mean. This looks more like an multipath configuration, or the lack of it. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-iscsi@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/open-iscsi -~--~~~~--~~--~--~---
Re: Logical Unit Not Ready, Initializing Cmd. Required
On Fri, May 16, 2008 at 08:28:43AM -0700, sinysee wrote: Hello, What do you mean by never gets available? Can you attach the full dmesg? Is it that the block device (/dev/sdX) that is unavailable or the multipath device (/dev/dm-XX)? I am looking to set up an initiator-target configuration with an FC- iSCSI bridge in between. As I understand the bridge should be transparently packaging an FC disk as an iSCSI target. After # iscsiadm -m node -p target portal -T target -l I get errors in dmesg and finally sd 6:0:0:1: Attached scsi disk sdb So the block disks show up fine. That means iSCSI works just fine. However # fdisk /dev/sdb Returns with Unable to read /dev/sdb Finally, # iscsiadm -m node -p target portal -T target -u returns. I am attaching the relevant part of the dmesg log. Hope it can clarify the problem You only see one disk? Either way, are you running multipath on your machine? If you are, what is the multipath -ll output? Dear Konrad, I find it strange that the host does not send Request Sense after receiving an error for Test Unit Ready. Is it a legit SCSI behaviour? The kernel intoragates the disks and that is what it gets. It reports to you these errors. There is no need to send a Request Sense command as the Test Unit Ready has already provided you the SCSI error, along with the ASC/ASCQ values, as you can see: sd 6:0:0:1: Device not ready: 6: Current: sense key: Not Ready Additional sense: Logical unit not ready, cause not reportable The next step for enterprise storages, such as the HP StorageWorks is to issue a START SCSI command - which is what you should see if you are using multipath. The path checker would start the LUN. --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups open-iscsi group. To post to this group, send email to open-iscsi@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/open-iscsi -~--~~~~--~~--~--~---