Hi Maged, thank you for your support. I've already read that book. The problem is that it uses pacemaker resources to manage iscsiTarget and iscsiLUNS. With this approach if I need a new target or a new lun I need to create a new resource in the cluster and this is an error prone task. My idea is to use a Filesystem as a resource where targetcli saves its configuration (/etc/target) when the global var auto_save_on_exit=true. With this method the configuration will be available to the new node in case of a failover. I'd like to know from the list if this is a good idea and if there is someone that uses this solution. Thank you, Marco 2017-03-27 20:18 GMT+02:00 Maged Mokhtar <mmokhtar@xxxxxxxxxxx>: > There are several online guides to manually build HA iSCSI using LIO and > DRBD. > However if you need something comprehensive I highly recommend this book: > Pro Linux High Availability Clustering > by Sander van Vugt > Apress > https://www.amazon.com/Pro-Linux-High-Availability-Clustering/dp/1484200802 > Large sections are devoted to exactly this. It is a real gem. > > /Maged > > > > From: Marco Marino > Sent: Monday, March 27, 2017 3:13 PM > To: Maged Mokhtar > Cc: target-devel@xxxxxxxxxxxxxxx > Subject: Re: High Availability solution for iscsi > > > > Hi Maged, > > thank you but.... I'm focused on drbd, ceph is not my favorite storage. > > > Marco > > > > 2017-03-27 13:48 GMT+02:00 Maged Mokhtar <mmokhtar@xxxxxxxxxxx>: > > > Hi Marco, > > you may be interested in PetaSAN: > www.petasan.org > It is an open source HA iSCSI solution using Ceph as its backstore. > > Cheers /maged > > -------------------------------------------------- > From: "Marco Marino" <marino.mrc@xxxxxxxxx> > Sent: Monday, March 27, 2017 12:30 PM > To: <target-devel@xxxxxxxxxxxxxxx> > Subject: High Availability solution for iscsi > > > Hi, I'm searching a solution for High availability of the target > service and backstore devices. My idea is to create a DRBD device > between 2 nodes and, on top of it, mount the folder /etc/target using > an ext4 filesystem. With this solution, when the active node goes > down, /etc/target and all the configuration will be remounted on the > other node. Obviously after the mount of the folder, the target > service will be started on the (new) active node. Furthermore, all > backstore devices are drbd resources so they can survive to a > failover. > I'd like to know if this is a good solution for HA. Is there some > other solution that I can use to ensure High Availability to iSCSI? > > Thank you, > Marco > > -- > To unsubscribe from this list: send the line "unsubscribe target-devel" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
