From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hannes Reinecke Date: Thu, 27 Oct 2005 11:50:33 +0000 Subject: Re: Fw: tape implementation question for SLES9 and persistent Binding Message-Id: <4360BF09.2050108@suse.de> List-Id: References: In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable To: linux-hotplug@vger.kernel.org Stefan Neff wrote: > Hi linux-hotplug-devels, >=20 > found on several internet web pages your email. I was looking for > implementation help for persistant binding of tape via udev-command under > SLES9 on a IBM pSerie with EMULEX LP10.000 >=20 Ah, yes. Up to know we hadn't considered that someone would need persistent device names for tapes. > The problem I / we is that we dont't realy understand all the theoretic > explamantion about the udev. All provided examples only talk about printer > and disk - not tape. > With disk and printeres it looks easy ... because for printers in the > "sysfs" is the s/n listed and for disk you can see the ID. With Tapes its > not the case - in the "sysfs" there is no S/N nor the WWNN or WWPN. So you > dont get information for a unique assignment from there. > Well, yes. But as (thankfully) most tapes are SCSI tapes anyway (or at least connected via a SCSI bridge) we can use the SCSI information. This is not really a problem. The main problem is that each tapes comes in different flavours. It has two modi (rewind and non-rewind) and several configurations, four of which are supported by Linux (that's what the 'a', 'l', or 'm' is about). To make matters even worse, all configurations (apart from the default one) have to be configured _explicitely_ otherwise they simply won't work. So, as to not confuse users even more I've got the following suggestion: tape/by-path/pci-0000:d8:01.0-scsi-1:0:0:0-st tape/by-path/pci-0000:d8:01.0-scsi-1:0:0:0-sta tape/by-path/pci-0000:d8:01.0-scsi-1:0:0:0-stl tape/by-path/pci-0000:d8:01.0-scsi-1:0:0:0-stm for the rewinding tapes; for non-rewinding add a 'n' in front of 'stX'. And if the tape supports SCSI EVPD pages we might even get a 'by-id' symlink like tape/by-id/200e09e600002d683-st ... If the tape doesn't support SCSI EVPD pages 0x80 or 0x83 then we'll have a hard time distinguishing between two tapes with the same make; hence the 'by-id' symlink won't be available there. D'accord? Cheers, Hannes --=20 Dr. Hannes Reinecke hare@suse.de SuSE Linux Products GmbH S390 & zSeries Maxfeldstra=DFe 5 +49 911 74053 688 90409 N=FCrnberg http://www.suse.de ------------------------------------------------------- This SF.Net email is sponsored by the JBoss Inc. Get Certified Today * Register for a JBoss Training Course Free Certification Exam for All Training Attendees Through End of 2005 Visit http://www.jboss.com/services/certification for more information _______________________________________________ Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net Linux-hotplug-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel