|
LTFS label or mount failed with IBM drives
Posted by Andre Kuehnemund, Last modified by Andre Kuehnemund on 01 August 2019 17:07
|
|
|
Customer setup: P5 version 5.6.4 RHEL/CentOS 7 server IBM LTO-7 tape drive(s) Customer tried to label a tape as LTFS media. The job failed. Excerpt from the P5 log: ... [31/Jul/2019:11:59:31][2404.7f03de5c0700][-conn:lexxsrv:gui:5] Notice: Job 23162 'Creating linear tape file system (LTFS)' queued to queue '0' ... [31/Jul/2019:12:00:31][2404.7f03dc2b6700][-thread:7f03dc2b6700-] Notice: sg231: mounting cartridge sg231: create a new label LTFS17085I Plugin: Loading "iokit" driver LTFS17085I Plugin: Loading "unified" iosched LTFS12164W Cannot detect lin_tape version LTFS12193E Old lin_tape version is detected. LTFS12012E Cannot open device: failed backend open call LTFS15009E Cannot open device '/dev/sg132' (-21701) LTFS15023I Formatting failed ... [31/Jul/2019:12:01:14][2404.7f03dc2b6700][-thread:7f03dc2b6700-] Notice: Other Job Error: Job error: 'Creating linear tape file system (LTFS)' (23162) Description : Creating linear tape file system (LTFS) Started At : Wed, 31. Jul 2019 11:59 Stopped At : Wed, 31. Jul 2019 12:01 Job Duration : 1 min 43 secs Job Result : error ------------ failed label for /dev/sg132 ... The issue was that IBM's 'lin_tape' driver included with RHEL/CentOS 7 is too old. IBM's LTFS implementation relies on newer versions of that driver in order for it to work. The customer downloaded an installed the latest version of'lin_tape' available from IBM. This solved the issue. As of this writing, the download link for lin_tape version 3.0.39 is as follows: https://www-945.ibm.com/support/fixcentral/swg/selectFixes?product=ibm%2FStorage_Tape%2FTape+device+drivers&fixids=lin_tape-3.0.39-rhel7.x86_64&source=dbluesearch&function=fixId&parent=Tape%20drivers%20and%20software The above link is obviously subject to change. However, a simple Google search for "lin_tape download" returned multiple hits for different versions. The minimum required version is 2.10. | |
 (0 vote(s)) |
Comments (0)