[Lustre-discuss] OSS Node Hard Disk Changing Procedure

[VIJESH EK]

Dear Sir,

 
We have a HPC setup with four OSS server(OSS1 to OSS4) and two MDS Nodes(MDS1 to MDS2)

OSS1 & OSS2 ,OSS3 & OSS4 clustered each other

In OSS4 logs i am getting the following   "unreadable" error continuously. So yesterday i runned fsck

and fixed some errors . Today the same type of error is comes again continuously.

Feb 26 06:24:43 oss4 smartd[9306]: Device: /dev/sda, 2 Currently unreadable (pending) sectors
Feb 26 06:54:43 oss4 smartd[9306]: Device: /dev/sda, 2 Currently unreadable (pending) sectors
Feb 26 07:24:43 oss4 smartd[9306]: Device: /dev/sda, 2 Currently unreadable (pending) sectors

/dev/sda is a local hard disk. So i desired to  change the local hard disk of the OSS4 server

I required a help from lustre team for the the easiest way to change the HDD in OSS Node.

The server is running now,so can  anybody suggest me the procedure to change the HDD in OSS servers ?

 

 

Rgds

 

VIJESH

 

 

[Artem Blagodarenko]

Hello,

I can't give exact script for changing HDD on OSS, but this document can help to 

understand OST deactivating and activating process, backup and restore procedures.

http://wiki.lustre.org/manual/LustreManual20_HTML/LustreMaintenance.html

Best regards,

Artem Blagodarenko.

 

_______________________________________________
Lustre-discuss mailing list
Lustre-...@lists.lustre.org
http://lists.lustre.org/mailman/listinfo/lustre-discuss

[Craig Tierney]

We have a need to use ACL's under lustre (server/client 1.8.7). I was
wondering if anyone is actually using them and if they work as documented?
Are there any known issues with ACLs?

Thanks,
Craig

[Mario David]

hi Craig

we still have lustre 1.8.6, but we (and several other sites) have been using acl's for a long time with no reported problem, i.e.
lustre is a backend FS for a thing called Storm (storage resource managers, a grid storage thing) exposed to external grid users
the way storm SW works it sets and modifies acl's according to whatever user or usage,

I didn't experiment or set "default" acl's, though I suppose they should also work OK

cheers
Mario

[Tina Friedrich]

Hi Craig,

assuming you're talking POSIX ACLs, they do, yes. We use ACLs (and
default ACLs) a lot.

One caveat being that we found that Lustre has a hard limit of 32 ACLs
on a file - this is, we believe, something that ext3 used to have (but
no longer has).

But if you don't need to set more than 32 ACLs on anything, not a problem.

Tina

--
Tina Friedrich, Computer Systems Administrator, Diamond Light Source Ltd
Diamond House, Harwell Science and Innovation Campus - 01235 77 8442

--
This e-mail and any attachments may contain confidential, copyright and or privileged material, and are for the use of the intended addressee only. If you are not the intended addressee or an authorised recipient of the addressee please notify us of receipt by returning the e-mail and do not use, copy, retain, distribute or disclose the information in or attached to the e-mail.
Any opinions expressed within this e-mail are those of the individual and not necessarily of Diamond Light Source Ltd.
Diamond Light Source Ltd. cannot guarantee that this e-mail or any attachments are free from viruses and we cannot accept liability for any damage which you may sustain as a result of software viruses which may be transmitted in or with the message.
Diamond Light Source Limited (company no. 4375679). Registered in England and Wales with its registered office at Diamond House, Harwell Science and Innovation Campus, Didcot, Oxfordshire, OX11 0DE, United Kingdom

[Kent Engström]

Mario David <da...@lip.pt> writes:
> hi Craig
>
> we still have lustre 1.8.6, but we (and several other sites) have been using acl's for a long time with no reported problem, i.e.
> lustre is a backend FS for a thing called Storm (storage resource managers, a grid storage thing) exposed to external grid users
> the way storm SW works it sets and modifies acl's according to whatever user or usage,
>
> I didn't experiment or set "default" acl's, though I suppose they should also work OK

There is an issue with default ACLs. See
http://jira.whamcloud.com/browse/LU-974,
"Lustre does not ignore umask when default ACL with mask is set"
(which contains proposed patches from Whamcloud for the issue).

--
Kent Engström, National Supercomputer Centre
ke...@nsc.liu.se, +46 13 28 4444

[Craig Tierney]

On 3/13/12 11:54 AM, Tina Friedrich wrote:
> Hi Craig,
>
> assuming you're talking POSIX ACLs, they do, yes. We use ACLs (and
> default ACLs) a lot.
>
> One caveat being that we found that Lustre has a hard limit of 32 ACLs
> on a file - this is, we believe, something that ext3 used to have (but
> no longer has).
>
> But if you don't need to set more than 32 ACLs on anything, not a problem.
>

Thanks for the response and the details. I do not believe that we need
that many. We are faking directory based quotas in lustre through the
use of sticky bit (g+s) and then running scripts to change group ownership
to fix problems. We just need to be able to tag certain files with a single
group ("restricted data") to block access to those without that group.
This should require only one ACL be set for the restricted data group.

Craig