| Summary: | NodeName expansion: One range must be at end requirement | ||
|---|---|---|---|
| Product: | Slurm | Reporter: | Kevin Buckley <kevin.buckley> |
| Component: | Configuration | Assignee: | Marshall Garey <marshall> |
| Status: | RESOLVED DUPLICATE | QA Contact: | |
| Severity: | 4 - Minor Issue | ||
| Priority: | --- | ||
| Version: | 22.05.2 | ||
| Hardware: | Cray Shasta | ||
| OS: | Linux | ||
| Site: | Pawsey | Alineos Sites: | --- |
| Atos/Eviden Sites: | --- | Confidential Site: | --- |
| Coreweave sites: | --- | Cray Sites: | --- |
| DS9 clusters: | --- | HPCnow Sites: | --- |
| HPE Sites: | --- | IBM Sites: | --- |
| NOAA SIte: | --- | OCF Sites: | --- |
| Recursion Pharma Sites: | --- | SFW Sites: | --- |
| SNIC sites: | --- | Linux Distro: | --- |
| Machine Name: | CLE Version: | ||
| Version Fixed: | Target Release: | --- | |
| DevPrio: | --- | Emory-Cloud Sites: | --- |
|
Description
Kevin Buckley
2023-07-05 01:59:17 MDT
(In reply to Kevin Buckley from comment #0) > Probably another one to mark as a dupe of 4138, although a > slightly different request. Ultimately the fix is the same. > Noting this requirement > > If one or more numeric expressions are included, one of them > must be at the end of the name (e.g. "unit[0-31]rack" is invalid), > but arbitrary names can always be used in a comma-separated list. > > Why is that. Because the hostlist code is designed to expect a prefix and then optionally node range(s) at the end, but does not support a suffix. Simply put: because it hasn't been implemented. > We, and I'd expect many other sites with Shasta, have the NodeAddr > names that we need to supply ending in the string "-nmn" which then > defeats the ability to use an expansion. > > I'll spare you full the config but here's a snippet > > NodeName=nid00[1000-1503] > NodeAddr=nid001000-nmn,nid001001-nmn,nid001002-nmn,nid001003-nmn,nid001004- > nmn,nid001005-nmn,nid001006-nmn,... > > It'd be really "nice", if we could just write > > NodeName=nid00[1000-1503] NodeAddr=nid00[1000-1503]-nmn We recognize this is a common problem. We are working to add support for a common suffix (for expanding hostlists, but perhaps not for recombining hostlists) in bug 16264. I'm closing this bug as a duplicate of bug 16264. (I think that bug 16264 is essentially a duplicate of 4138; however, I think when that bug was opened they did not realize that 4138 exists.) *** This ticket has been marked as a duplicate of ticket 16264 *** |