Sponsored Content
Operating Systems Linux EMC, PowerPath and issue on using LUN Post 302582272 by Zarnick on Thursday 15th of December 2011 09:36:53 AM
Old 12-15-2011
Simply put, we don't have support for this storage (at least for that, only for incidents), at least that's what I've being told Smilie .
 

10 More Discussions You Might Find Interesting

1. Filesystems, Disks and Memory

Can Veritas DMP & EMC PowerPath coexist?

We currently have a Solaris box connected to a Clariion storage system that is utilising DMP for path failover. I would prefer to use EMC's Powerpath and was wondering if the two can coexist? Basically, I am struggling to find any documentation on the subject and was wonder if anyone can give me... (2 Replies)
Discussion started by: aaron2k
2 Replies

2. Solaris

I have LUN ID, how to find disk relate to that LUN ID?

I have a list of LUN ID, my task is to find if disk has been added or not. How do I do that? I have been searching the forum and not able to find answer. thanks (4 Replies)
Discussion started by: uuontario
4 Replies

3. Red Hat

Configure EMC Powerpath?

Hi , I have a redhat 5.3 server which has 2 vg.. one is rootvg in local harddisk and another one is applicationvg in SAN.. When I reboot the server , EMC powerpath driver is not starting up automatically. Hence applicationvg is not mounting properly. Therefore I need to unmount it manually and... (4 Replies)
Discussion started by: Makri
4 Replies

4. UNIX for Advanced & Expert Users

Removing EMC LUN from Solaris 10 host without a reboot

I have seen a few threads / posts on the forum's but nothing specific to what Iam after. So Iam hoping you can help. Issue : Iam hoping to remove unused emc luns from my solaris servers , T2000 (5.10 Generic_142900-02). I have a list of the process I have followed for the removal of the luns.... (3 Replies)
Discussion started by: furby
3 Replies

5. Solaris

Solaris/vxvm/EMC Lun configuration

Hello all, i try to allocate the same LUN to two server (or more in the future) i use solaris 10, vxvm (vxfs) for data and solaris zones and EMC DMX-4, i try to migrate solaris zones between servers in case of problem. and this is what i want to do - affect LUN to srv00124 and srv10155 -... (5 Replies)
Discussion started by: mat_solaris
5 Replies

6. Solaris

One emc powerpath failed

It seems like I lost one path on my Solaris-11 box. But I want to make sure before going to Storage team, if issue is from OS side or Storage side. Storage team is able to see that only one wwwn is looged in their switch. I am not at server's physical location. What does below output says ? I... (0 Replies)
Discussion started by: solaris_1977
0 Replies

7. Solaris

Emc powerpath device & zfs query

Hi, We're trying out a SAN migration from HP EVA to EMC VMAX, and run into a bit of an issue with powerpath and zfs. The method we're currently using to migrate is to export the HP EVA luns from our sun server, replicate using SAN based method, and then present the new luns to our Sun server... (7 Replies)
Discussion started by: badoshi
7 Replies

8. Solaris

Solaris OS having issue accessing LUN from Qlogic swich

Hi Community, I am facing one issue related storage accessing from solaris 10 machine(M5000). it was working fine then we shifted to another site. the only change on that site is FC switch. The issue i am facing is from solaris machine i can able to see the LUN, it;s only a test lun of 30... (17 Replies)
Discussion started by: bentech4u
17 Replies

9. AIX

AIX Migration issue with EMC ODM sets

Hi Experts , I want to start migrating our AIX 6.1 to AIX 7.1 . I am planning to use alt_disk_migration . Chris gibson has awesome documentation in the internet. However I am running into an issue with EMC odm filesets . So my current OS is AIX 6.1. and I have this : lslpp -l | grep EMC ... (7 Replies)
Discussion started by: JME2015
7 Replies

10. AIX

AIX help -reconnect EMC lun

we have an old AIX system with important data on an EMC LUN. the AIX server crashed and we had to rebuild it with the old AIX os 5.8. how do we reconnect the EMC LUN without losing the data? (15 Replies)
Discussion started by: jhudson12
15 Replies
SG_LUNS(8)							     SG3_UTILS								SG_LUNS(8)

NAME
sg_luns - send SCSI REPORT LUNS command or decode given LUN SYNOPSIS
sg_luns [--decode] [--help] [--hex] [--linux] [--maxlen=LEN] [--quiet] [--raw] [--select=SR] [--verbose] [--version] DEVICE sg_luns --test=ALUN [--hex] [--verbose] DESCRIPTION
In the first form shown in the SYNOPSIS this utility sends the SCSI REPORT LUNS command to the DEVICE and outputs the response. The response should be a list of LUNs ("a LUN inventory") for the I_T nexus associated with the DEVICE. Roughly speaking that is all LUNs that share the target device that the REPORT LUNS command is sent through. In the SPC-3 and SPC-4 SCSI standards support for the REPORT LUNS command is mandatory. When the --test=ALUN option is given (the second form in the SYNOPSIS), then the ALUN value is decoded as outlined in SAM-3, SAM-4 and SAM-5 (revision 13, section 4.7) . Where required below the first form shown in the SYNOPSIS is called "device mode" and the second form is called "test mode". OPTIONS
Arguments to long options are mandatory for short options as well. -d, --decode decode LUNs into their component parts, as described in the LUN section of SAM-3, SAM-4 and SAM-5. -h, --help output the usage message then exit. -H, --hex [device mode] when given once this utility will output the SCSI response (i.e. the data-out buffer) to the REPORT LUNS command in ASCII hex then exit. When given twice it causes --decode to output component fields in hex rather than decimal. -H, --hex [test mode] when this option is given, then decoded component fields of ALUN are output in hex. -l, --linux this option is only available in Linux. After the T10 representation of each 64 bit LUN (in 16 hexadecimal digits), if this option is given then to the right, in square brackets, is the Linux LUN integer in decimal. If the --hex option is given twice (e.g. -HH) as well then the Linux LUN integer is output in hexadecimal. -m, --maxlen=LEN where LEN is the (maximum) response length in bytes. It is placed in the cdb's "allocation length" field. If not given (or LEN is zero) then 8192 is used. The maximum allowed value of LEN is 1048576. -q, --quiet output only the ASCII hex rendering of each report LUN, one per line. Without the --quiet option, there is header information printed before the LUN listing. -r, --raw output the SCSI response (i.e. the data-out buffer) in binary (to stdout). -s, --select=SR this option sets the SELECT REPORT field (SR) in the SCSI REPORT LUNS command. The default value is 0. For detailed information see the REPORT LUNS command in SPC (most recent is SPC-4 revision 36e in section 6.33). To simplify, for the I_T nexus associated with the DEVICE, the meanings of the SR values defined to date for SPC-4 are: 0 : all LUNs excluding well known logical units 1 : well known logical units 2 : all LUNs Values between 0xf8 and 0xff (inclusive) are vendor specific (SPC-4 rev 36e), other values greater than 2 are reserved. This utility will accept any value between 0 and 255 (0xff) for SR . -t, --test=ALUN ALUN is assumed to be a hexadecimal number in ASCII hex or the letter 'L' followed by a decimal number (see below). The hexadecimal number can be up to 64 bits in size (i.e. 16 hexadecimal digits) and is padded to the right if less than 16 hexadecimal digits are given (e.g. --test=0122003a represents T10 LUN 0122003a00000000). ALUN may be prefixed by '0x' or '0X' (e.g. the previous example could have been --test=0x0122003a). ALUN may also be given with spaces or tabs between each byte (or other grouping) but then ALUN would need to be surrounded by single or double quotes. In the decimal number case (i.e. following a 'L') that number is assumed to be a Linux "word flipped" LUN which is converted into a T10 LUN representation and printed. In both cases the number is interpreted as a LUN and decoded as if the --decode option had been given. Also when ALUN is a hexadecimal number it can have a trailing 'L' in which case the corresponding Linux "word flipped" LUN value is output. The LUN is decoded in all cases. -v, --verbose increase the level of verbosity, (i.e. debug output). -V, --version print the version string and then exit. NOTES
The SCSI REPORT LUNS command is important for Logical Unit (LU) discovery. After a target device is discovered (usually via some transport specific mechanism), a REPORT LUNS command should either be sent to LUN 0 (which is Peripheral device addressing method with bus_id=0 and target/lun=0) or to the REPORT LUNS well known LUN (i.e. 0xc101000000000000). SAM-5 requires that one of these responds with an inventory of LUNS that are contained in this target device. In test mode, if the --hex option is given once then in the decoded output, some of the component fields are printed in hex with leading zeros. The leading zeros are to indicate the size of the component field. For example: in the Peripheral device addressing method (16 bits overall), the bus ID is 6 bits wide and the target/LUN field is 8 bits wide; so both are shown with two hex digits (e.g. bus_id=0x02, tar- get=0x3a). EXAMPLES
Typically by the time user space programs get to run, SCSI LUs have been discovered. In Linux the lsscsi utility lists the LUs that are currently present. The LUN of a device (LU) is the fourth element in the tuple at the beginning of each line. Below we see a target (or "I_T Nexus": "6:0:0") has two LUNS: 1 and 49409. If 49409 is converted into T10 LUN format it is 0xc101000000000000 which is the REPORT LUNS well known LUN. # lsscsi -g [6:0:0:1] disk Linux scsi_debug 0004 /dev/sdb /dev/sg1 [6:0:0:49409]wlun Linux scsi_debug 0004 - /dev/sg2 We could send a REPORT LUNS command to either /dev/sdb, /dev/sg1 or /dev/sg2 and get the same result. Below we use /dev/sg1 : # sg_luns /dev/sg1 Lun list length = 8 which imples 1 lun entry Report luns [select_report=0x0]: 0001000000000000 That is a bit noisy so cut down the clutter with --quiet: # sg_luns -q /dev/sg1 0001000000000000 Now decode that LUN into its component parts: # sg_luns -d -q /dev/sg1 0001000000000000 Peripheral device addressing: lun=1 Would like to see how wide that component LUN field is: # sg_luns -d -q -HH /dev/sg1 0001000000000000 Peripheral device addressing: lun=0x01 So it is 8 bits wide (actually between 5 and 8 bits wide, inclusive). Now use --select=1 to find out if there are any well known LUNs: # sg_luns -q -s 1 /dev/sg1 c101000000000000 So how many LUNs do we have all together (associated with the current I_T Nexus): # sg_luns -q -s 2 /dev/sg1 0001000000000000 c101000000000000 # sg_luns -q -s 2 -d /dev/sg1 0001000000000000 Peripheral device addressing: lun=1 c101000000000000 REPORT LUNS well known logical unit The following example uses the --linux option and is not available in other operating systems. The extra number in square brackets is the Linux version of T10 LUN shown at the start of the line. # sg_luns -q -s 2 -l /dev/sg1 0001000000000000 [1] c101000000000000 [49409] Now we use the --test= option to decode LUNS input on the command line (rather than send a REPORT LUNS command and act on the response): # sg_luns --test=0001000000000000 Decoded LUN: Peripheral device addressing: lun=1 # sg_luns --test="c1 01" Decoded LUN: REPORT LUNS well known logical unit # sg_luns -t 0x023a004b -H Decoded LUN: Peripheral device addressing: bus_id=0x02, target=0x3a >>Second level addressing: Peripheral device addressing: lun=0x4b The next example is Linux specific as we try to find out what the Linux LUN 49409 translates to in the T10 world: # sg_luns --test=L49409 64 bit LUN in T10 preferred (hex) format: c1 01 00 00 00 00 00 00 Decoded LUN: REPORT LUNS well known logical unit And the mapping between T10 and Linux LUN representations can be done the other way: # sg_luns -t c101L Linux 'word flipped' integer LUN representation: 49409 Decoded LUN: REPORT LUNS well known logical unit EXIT STATUS
The exit status of sg_luns is 0 when it is successful. Otherwise see the sg3_utils(8) man page. AUTHORS
Written by Douglas Gilbert. REPORTING BUGS
Report bugs to <dgilbert at interlog dot com>. COPYRIGHT
Copyright (C) 2004-2013 Douglas Gilbert This software is distributed under a FreeBSD license. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PUR- POSE. SEE ALSO
sg_inq(8) sg3_utils-1.36 May 2013 SG_LUNS(8)
All times are GMT -4. The time now is 01:53 PM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy