Antw: Re: Q: RAID1 with open-iscsi

2010-10-07 Thread Ulrich Windl
>>> Mike Christie schrieb am 08.10.2010 um 05:07 in Nachricht <4cae8b0a.9050...@cs.wisc.edu>: > On 10/07/2010 08:43 AM, Ulrich Windl wrote: > > Hi, > > > > a question: From what I've read, md won't work with open-iscsi (in SLES10). > What options do exist to have a RAID1 setup using two different

Antw: Re: Q: RAID1 with open-iscsi

2010-10-07 Thread Ulrich Windl
>>> Ben Turner schrieb am 07.10.2010 um 15:54 in Nachricht <1465692532.1411741286459697916.javamail.r...@zmail07.collab.prod.int.phx2.redha .com>: > What are you looking for more throughput or redundancy? You could bond NICs > or use 10GB nics for throughput and/or use DM Multipath for redunda

Re: Q: RAID1 with open-iscsi

2010-10-07 Thread Mike Christie
On 10/07/2010 08:43 AM, Ulrich Windl wrote: Hi, a question: From what I've read, md won't work with open-iscsi (in SLES10). What options do exist to have a RAID1 setup using two different storage systems? I only have experience with md and local disks. I have never heard that. Why doesn't i

Re: Strange iSCSI state ([iscsi_unbind_20], not logged in, hanging)

2010-10-07 Thread Mike Christie
Hey Hannes, Do you have the iscsi kernel source for kernel-smp-2.6.16.60-0.66.1? In one of your kernels did you add a iscsi_unbind thread? On 10/07/2010 05:32 PM, Mike Christie wrote: On 10/07/2010 05:17 AM, Ulrich Windl wrote: Warning: bad ps syntax, perhaps a bogus '-'? See http://procps.

Re: Strange iSCSI state ([iscsi_unbind_20], not logged in, hanging)

2010-10-07 Thread Mike Christie
On 10/07/2010 05:17 AM, Ulrich Windl wrote: Warning: bad ps syntax, perhaps a bogus '-'? See http://procps.sf.net/faq.html 3522 ?S< 0:00 [iscsi_eh] 3543 ?Ss 0:00 /sbin/iscsid -c /etc/iscsi/iscsid.conf -p /var/run/iscsi.pid 3544 ?S . Do you have the so

Re: [PATCH RFC/RFT] add iscsi dev loss timeout and device removal

2010-10-07 Thread Mike Christie
On 10/07/2010 02:02 AM, Hannes Reinecke wrote: Mike Christie wrote: Hey, At storage summit this year, it was decided iscsi needs to catch up with the other iscsi drivers, and add something like FC's dev_loss_tmo. When this timeout expires, it will cause the iscsi layer to remove the scsi device

Res: Q: RAID1 with open-iscsi

2010-10-07 Thread Marcos
Hi Ulrich, you ca setup a mirror with drbd which can work with iscsi. Marcos - Mensagem original De: Ulrich Windl Para: open-iscsi Enviadas: Quinta-feira, 7 deOutubrode 2010 10:43:03 Assunto: Q: RAID1 with open-iscsi Hi, a question: From what I've read, md won't work with open-is

Re: How can I dedicate iSCSI traffic to only 1 nic.

2010-10-07 Thread Ben Turner
Have you tried specifying -I, --interface[iface] when you log into the target? Not 100% but I think that will do what you want. -b - "jimmy" wrote: > Hi all, > > I have a server with two NICs on the same subnet. How can I force > iSCSI traffic to only use one of these? > I see the read

Strange iSCSI state ([iscsi_unbind_20], not logged in, hanging)

2010-10-07 Thread Ulrich Windl
Hi, this was detected on SLES10 SP3 (open-iscsi-2.0.868-0.6.11, kernel-smp-2.6.16.60-0.66.1) 4-core x86_64 (Opteron) machine: One of two iSCSI targets said the initiator is not logged in. Actually I see no attempts on the initiator for log ins. When trying a "iscsiadm -m node -L all", the comm

Re: Q: RAID1 with open-iscsi

2010-10-07 Thread Ben Turner
What are you looking for more throughput or redundancy? You could bond NICs or use 10GB nics for throughput and/or use DM Multipath for redundancy. Each NIC for single NICs or bond for bonded interfaces should have its own iface configured, check the man page for iscsiadm for details ther. -b

Q: RAID1 with open-iscsi

2010-10-07 Thread Ulrich Windl
Hi, a question: From what I've read, md won't work with open-iscsi (in SLES10). What options do exist to have a RAID1 setup using two different storage systems? I only have experience with md and local disks. Regards, Ulrich -- You received this message because you are subscribed to the Goog

Re: [PATCH RFC/RFT] add iscsi dev loss timeout and device removal

2010-10-07 Thread Hannes Reinecke
Mike Christie wrote: > Hey, > > At storage summit this year, it was decided iscsi needs to catch up with > the other iscsi drivers, and add something like FC's dev_loss_tmo. When > this timeout expires, it will cause the iscsi layer to remove the scsi > devices and fail any IO that was queued. Upp