Sorry, I'm posting this question again as my original post go abducted by a different thread due to stupid way I replied to that thread and changed the subject as a new post. My apologies if there's any confusion. On 14/04/15 10:05, Tom Robinson wrote: > # cat /etc/redhat-release > CentOS Linux release 7.1.1503 (Core) > # uname -r > 3.10.0-123.20.1.el7.x86_64 > > Hi, > > We use iSCSI over a 10G Ethernet Adapter and SRP over an Infiniband adapter to provide multipathing > to our storage: > > # lspci | grep 10-Gigabit > 81:00.0 Ethernet controller: Intel Corporation 82599ES 10-Gigabit SFI/SFP+ Network Connection (rev 01) > 81:00.1 Ethernet controller: Intel Corporation 82599ES 10-Gigabit SFI/SFP+ Network Connection (rev 01) > # lspci | grep Mellanox > 06:00.0 InfiniBand: Mellanox Technologies MT26428 [ConnectX VPI PCIe 2.0 5GT/s - IB QDR / 10GigE] > (rev b0) > > Originally the HBAs came up as host0, host1 and host2. After recent reboots we get host1, host2 and > host3: > > # ls -l /sys/class/scsi_host/host* > lrwxrwxrwx. 1 root root 0 Apr 9 19:10 /sys/class/scsi_host/host1 -> > ../../devices/platform/host1/scsi_host/host1 > lrwxrwxrwx. 1 root root 0 Apr 9 08:12 /sys/class/scsi_host/host2 -> > ../../devices/pci0000:00/0000:00:03.2/0000:06:00.0/host2/scsi_host/host2 > lrwxrwxrwx. 1 root root 0 Apr 9 08:12 /sys/class/scsi_host/host3 -> > ../../devices/pci0000:00/0000:00:03.2/0000:06:00.0/host3/scsi_host/host3 > > I've observed that at different times during configuration of this host we have had varying > combinations of the mapping of HBA to an enumerated /sys node. e.g.: > > host0 > host2 > host3 > > or > > host1 > host3 > host4 > > other combinations, etc. > > This messes up our multipath filtering for weightedpath priority settings. From /etc/multipath.conf: > > defaults { > user_friendly_names no > find_multipaths yes > prio weightedpath > prio_args "hbtl 0:.:.:.* 1 [1-9]:.:.:.* 10" > path_grouping_policy group_by_prio > path_selector "service-time 0" > failback immediate > polling_interval 5 > no_path_retry 12 > } > > The setting for prio_args fails to set the correct path map topology and priority against an HBA as > I can't predict what HBA is what. Consistently iSCSI claims the lowest number but I can't say if > that's going to be 0 or 1 a this stage. > > Does anyone know how to configure HBAs to stay on a fixed enumeration at boot? > > Kind regards, > Tom > > > > _______________________________________________ > CentOS mailing list > CentOS@xxxxxxxxxx > http://lists.centos.org/mailman/listinfo/centos
Attachment:
signature.asc
Description: OpenPGP digital signature
_______________________________________________ CentOS mailing list CentOS@xxxxxxxxxx http://lists.centos.org/mailman/listinfo/centos