-
Notifications
You must be signed in to change notification settings - Fork 1.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
vdev_id not properly detecting multipath disk configuration #11095
Comments
Seems I'm running the test-version of vdev_id wrong, so here's the right way for the same devices:
It looks like something about the ID stuff is breaking down and the symlinks are clobbering each other. I will investigate the identifier config. |
Changing the |
Adding |
System information
Describe the problem you're observing
I have a system with 2x LSI SAS3008 PCI-Express Fusion-MPT SAS-3 connecting to two IO controllers on a Supermicro 90-disk JBOD: https://www.supermicro.com/en/products/chassis/4U/946/SC946ED-R2KJBOD in a crossover dual-channel config
This is intended to provide cross-failover protection in case of card or IO controller failure. I currently have this configured with "multibus" in multipathd, and it is properly detecting and setting up 90 aggregated paths:
Snip of multipath -ll, some addresses greeked
When attempting to use
vdev_id.conf
using a standard multipath config:I get the following
/dev/disk/by-vdev
entries:Note that this is missing many entries, currently totaling only 40.
When testing vdev_id for bugs, it does appear to throw an error:
It returns the same error for mapper entries that do not get vdevs generated:
I'm unclear from the documentation what a number of the terms are for the vdev_id.conf file in the context of real-world hardware configurations (how do I determine number of phys, how do I know if switch or direct, is a HBA port the actual physical wire, what does "CHANNEL NAME" refer to?), so I'm having much difficulty guessing as to what settings might need to be changed.
Finally, I offer the output of sas_discover from https://github.com/stanford-rc/sasutils which appears to have mapped the entire SAS topology, it does look a bit odd compared to what I expect to see, this may be be due to the internal wiring of the JBOD
It looks like internally the IO controllers each have 1 expander going to 12 disks, and cascading 2 more expanders with 39 disks each, a very odd config, but each of the 90 disks is exposed properly on each physical wiring path, as expected.
The text was updated successfully, but these errors were encountered: