[Halld-offline] Fwd: Notes for systems mounting Lustre on Ethernet through the NFS Gateways

[Mark Ito]

from Bryan Hess...



-------- Forwarded Message --------
Subject: 	Notes for systems mounting Lustre on Ethernet through the NFS 
Gateways
Date: 	Wed, 7 Aug 2019 16:08:19 -0400
From: 	Bryan Hess <bhess at jlab.org>
To: 	Brad Sawatzky <brads at jlab.org>, Ole Hansen <ole at jlab.org>, Mark Ito 
<marki at jlab.org>, Nathan Baltzell <baltzell at jlab.org>, Sergey Boyarinov 
<boiarino at jlab.org>, David Lawrence <davidl at jlab.org>
CC: 	sciadm at jlab.org <sciadm at jlab.org>, Graham Heyes <heyes at jlab.org>



Hall or desktop system that are access /lustre/expphy/cache over 
Ethernet using the NFS gateways may need changes after tomorrow's Lustre 
maintenance

*Automounter*:

As of this writing there is a new entry for NIS auto.v.bb map that 
provides lustre as a mount point via scifsgw10a (also has the alias 
scifsgw) which is already in place. This should make the transition 
straightforward in most cases:

$ showmount -e scifsgw10a
Export list for scifsgw10a:
/lustre19 @workmount,cni-adm1.jlab.org
/lustre2  @workmount,cni-adm1.jlab.org

These changes have been made for the scidaqgw10a/b/c/d/e/f systems.


*Non-Automounter (/etc/fstab):*

If direct mounts for cache are done in /etc/fstab then they will need to 
make the needed adjustments for each case changing any of there local links.

Also need to make clear that both old and new lustre will have to 
co-exist since volatile resides on old and cache resides on new.

In summary: Systems need to have both /v/lustre2 and /v/lustre19 after 
tomorrow morning.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/halld-offline/attachments/20190807/b2ff5916/attachment-0001.html>