toasters

toasters@lists.teaparty.net

1 participants
13528 discussions

Recommendations on preserving a filers data?
by Arnold de Leon 22 Jun '15

22 Jun '15

What are the recommendations for the "best" way to preserve a copy data on a filer for future use? Background: A company has ceased operations and wants to preserve copies of its data "just in case someone wants it in the future". All the interesting data has been copied to the a single "archive" filer. There are about eight volumes in using about 50TB (with dedupe but no compression). This will be backup copy #1. The other filers that previously held the data are being sold and scheduled to be wiped. The data is a mixed of virtual machines and file shares. Now I'm trying to see if if a way to make another copy of the data for safe keeping at a reasonable cost. Speed of restore would not be high the feature list of this secondary archive. One of the first thoughts simply dump the data onto an AWS volume for "cold storage". Sane? Tips/tricks? Data presumably gets scrubbed to prevent bit rot. Conveniently available just about anywhere. Has a recurring cost but avoids some upfront cost. Second thought is find a relatively inexpensive array of disks and dump the data to that. Most likely benefit is fastest to write to. Maybe cheapest but in the end all the data is on non-spinning disks without verification. Tapes? I don't like tapes. Don't know if a working tape system available but if there was is this a good case for it. Tape static storage better than disk? Thanks. arnold

7 8

Clustered ONTAP 8.2 routing
by Legato, John (NIH/NHLBI) [E] 17 Jun '15

17 Jun '15

Hello, We've recently acquired a 2 node FAS8020 cluster running Clustered DataONTAP 8.2. We inherited a rather unusual/silly network topology, in which we have two subnets (a publicly routable one and an RFC1918 one for internal use only) running on the same VLAN. In other words, there's no router between these two subnets. In Linux, we're able to get boxes on the publicly routable subnet to communicate with the internal subnet (10.1.0.0/16) as follows: route add -net 10.1.0.0 netmask 255.255.0.0 dev eth0 This command tells the Linux box that it can send packets to 10.1.0.0/16 hosts out eth0 without going through an intervening router. However, we've yet to find a way to do this with Cluster DataONTAP. We've played around with routing groups and routes, but we cannot figure out a way to tell Clustered DataONTAP to route through an interface. Does anyone know if this is possible? NetApp has advised us to set up a router between our two subnets. We share NetApp's concern over fail over We'll set up a router if we can't figure out any other way to make this work. Thanks John

2 1

issue with Ontap upgrade 7mode system
by Sayla, Mustafa 15 Jun '15

15 Jun '15

I recently upgraded one of my 7mode controller from 8.1P1 to 8.2.3P3 and ran into issue during the second controller upgrade. I installed the upgrade version on both the controller than performed the cf takeover from controller B to controller A and it went fine. I then waited about 10 minutes and did the cf giveback and that also went fine. after waiting another 10 minutes I tried the cf takeover from controller A to controller B and I got the version mismatch error. I then issued cf takeover -n and at the time controller A did not took over controller B but just rebooted controller B. This caused a downtime with our vmware farm as servers lost connection to the data store on controller B. This is with NFS protocol. I only see this error in the logs Fri May 29 15:53:59 CDT [rddata1b: cf_takeover: kern.rc.msg:notice]: vlan: interface creation failed due to internal errors Partner vlan create command failed This explains why the takeover did not work but does not explain why this happened in the first place. I have worked with Netapp and validated that the rc file is correct and we have done failover and failback in the past without any issue. I also did the failover and failback after this to rule out configuration issue but could not recreate it. I need to upgrade 3 more HA pairs and I want to know if anyone has seen this issue before and does this have anything to do with the upgrade. Thank you Mustafa Sayla Visit us on the Web at mesirowfinancial.com This communication may contain privileged and/or confidential information. It is intended solely for the use of the addressee. If you are not the intended recipient, you are strictly prohibited from disclosing, copying, distributing or using any of this information. If you received this communication in error, please contact the sender immediately and destroy the material in its entirety, whether electronic or hard copy. Confidential, proprietary or time-sensitive communications should not be transmitted via the Internet, as there can be no assurance of actual or timely delivery, receipt and/or confidentiality. This is not an offer, or solicitation of any offer to buy or sell any security, investment or other product.

2 2

cifs.smb2.client.enable
by Ray Van Dolson 15 Jun '15

15 Jun '15

cifs.smb2.client.enable This option enables SMB 2.0 client capability on the Filer. When this option is enabled, Filer-initiated connections to Windows servers use the SMB 2.0 protocol. If the Windows server does not support the SMB 2.0 protocol, the Filer uses SMB 1.0. If a session was established over SMB 2.0 and then this option is disabled, existing sessions are not terminated. The Filer continues to use SMB 2.0 for the existing sessions; new sessions do not use SMB 2.0. Probably missing something obvious -- but would someone describe to me a scenario in which the NetApp would act as a CIFS client? Thanks, Ray

3 2

NetBIOS Cleartext Password Access
by Fletcher Cocquyt 15 Jun '15

15 Jun '15

We ran a security scan (using Qualys) and it reports: NetBIOS Cleartext Password Access What’s the recommended way to remedy this? We run 99% NFS, is the best solution to turn off CIFS? thanks

2 1

unified manager for CDOT 8.3
by Sayla, Mustafa 13 Jun '15

13 Jun '15

We recently installed our first CDOT system with 8.3P1 and now I want to setup the monitoring and configuration environment for it. I have few questions * Do I need a separate OnCommand Server for CDOT or I can use my existing server that I use to manage my 7mode systems. My current system is using OnCommand 5.1. * We are on Vsphere 5.0 and per admin guide it seems like I cannot use Unified Manager 6.2 so can I use Unified Manager 6.1 with CDOT 8.3P1 * Is performance manager part of unified Manager or it is a separate appliance. What I want to know that if I download and install the unified manager appliance do I get performance manager as part of it or I have to download performance manager appliance separately and install it as a separate VM Thank you Mustafa Sayla Visit us on the Web at mesirowfinancial.com This communication may contain privileged and/or confidential information. It is intended solely for the use of the addressee. If you are not the intended recipient, you are strictly prohibited from disclosing, copying, distributing or using any of this information. If you received this communication in error, please contact the sender immediately and destroy the material in its entirety, whether electronic or hard copy. Confidential, proprietary or time-sensitive communications should not be transmitted via the Internet, as there can be no assurance of actual or timely delivery, receipt and/or confidentiality. This is not an offer, or solicitation of any offer to buy or sell any security, investment or other product.

3 2

Mystery quota on qtree?
by Jeff Cleverley 12 Jun '15

12 Jun '15

Greetings, I have a 8.1.2 7-mode cluster and was asked to adjust a couple of quotas. The problem is there doesn't appear to be a quota set on the trees. If I cd to the top of the volume and do a "df -h .", I get the volume totals. When I go into any of the sub-directories, it shows me different numbers for size allocated and used. The free is consistent at all levels but not the size and used. I've looked at the /etc/quotas, quota report, and viewed it through OnCommand. Nothing shows a quota limit. I've also run pwd-P and the df -h . both show I'm in the same file system and have not followed a symlink. The space reservation is volume and guaranteed, and snap reserve is at 0% utilized and has not grown beyond it's reserve. Below is the best I can give for details. Thanks, Jeff [root@admin storage1]# df -h . ## top level volume Filesystem Size Used Avail Use% Mounted on filer2:/vol/storage1 2.4T 1.1T 1.3T 45% /net/filer2/vol/storage1 [root@admin storage1]# cd tools [root@admin tools]# df -h . Filesystem Size Used Avail Use% Mounted on filer2:/vol/storage1 1.4T 71G 1.3T 6% /net/filer2/vol/storage1 [root@admin tools]# cd ../home [root@admin home]# df -h . Filesystem Size Used Avail Use% Mounted on filer2:/vol/storage1/home 1.5T 118G 1.3T 9% /net/filer2/vol/storage1/home [root@admin home]# cd .. [root@admin storage1]# df -h win_shares Filesystem Size Used Avail Use% Mounted on filer2:/vol/storage1 2.2T 843G 1.3T 39% /net/filer2/vol/storage1 [root@admin etc]# cat quotas #Auto-generated by setup Tue Jun 4 03:55:56 GMT 2013 /vol/aus_ntshares/ashares tree - - - - - /vol/storage1/apps tree - - - - - /vol/storage1/home tree - - - - - /vol/storage1/tools tree - - - - - /vol/storage1/win_shares tree - - - - - [root@admin storage1]# ssh filer2 quota report K-Bytes Files Type ID Volume Tree Used Limit Used Limit Quota Specifier ----- -------- -------- -------- --------- --------- ------- ------- --------------- tree 1 storage1 apps 10421344 - 46712 - /vol/storage1/apps tree 2 storage1 home 123265872 - 796745 - /vol/storage1/home tree 3 storage1 tools 73659616 - 773186 - /vol/storage1/tools -- Jeff Cleverley IT Engineer 4380 Ziegler Road Fort Collins, Colorado 80525 970-288-4611

4 4

Re: Disk auto assign
by Edward Rolison 12 Jun '15

12 Jun '15

Spot on - that's done the trick. Had '0a.01.0' with a foreign owner ID hiding in the list somewhere. I can only assume as part of a reboot/rebuild/pre-ship-test or something. Reassigned it to 'unowned' and all good. (well, a 'foreign aggr0(1)' showed up). On 11 June 2015 at 13:15, tmac <tmacmd(a)gmail.com> wrote: > SO, look at the output of disk show -a > The -n shows all non-owned disks. > > Look for disks not owned by either of your two nodes now. > If there are any disks owned by another party....ugh, little bit of work > involved. > > If not, do your best to verify which disks on which shelves are owned and > where. > > Sometimes it is helpful to run through the disks one by one, if possible, > use a sorting routine to put the disks in order to make looking at the list > easier. > > Let us know! > > > --tmac > > *Tim McCarthy* > *Principal Consultant* > > On Thu, Jun 11, 2015 at 8:12 AM, tmac <tmacmd(a)gmail.com> wrote: > >> any chance you can copy to a flash drive or the like? or is it more >> secure than that ;) >> >> >> --tmac >> >> *Tim McCarthy* >> *Principal Consultant* >> >> >> >> On Thu, Jun 11, 2015 at 8:11 AM, Edward Rolison <ed.rolison(a)gmail.com> >> wrote: >> >>> Sadly I can't - this is a system on an isolated network. I can re-type >>> parts of it, but that's probably not as useful. :/ >>> >>> On 11 June 2015 at 12:42, tmac <tmacmd(a)gmail.com> wrote: >>> >>>> Hey Edward, just for kicks, how about posting the "disk show -a" for us. >>>> >>>> Thanks >>>> >>>> --tmac >>>> >>>> *Tim McCarthy* >>>> *Principal Consultant* >>>> >>>> >>>> On Thu, Jun 11, 2015 at 7:26 AM, Edward Rolison <ed.rolison(a)gmail.com> >>>> wrote: >>>> >>>>> New disks are all listed in 'disk show -n'. I think that means they're >>>>> not assigned. (And a disk I've deliberately unassigned is also not being >>>>> reassigned). >>>>> >>>>> On 11 June 2015 at 12:12, Borzenkov, Andrei < >>>>> andrei.borzenkov(a)ts.fujitsu.com> wrote: >>>>> >>>>>> Are you sure new disks are unowned? It is not uncommon to get >>>>>> repurposed hardware. >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> *From:* Edward Rolison [mailto:ed.rolison@gmail.com] >>>>>> *Sent:* Thursday, June 11, 2015 2:09 PM >>>>>> *To:* Borzenkov, Andrei >>>>>> *Cc:* toasters(a)teaparty.net >>>>>> *Subject:* Re: Disk auto assign >>>>>> >>>>>> >>>>>> >>>>>> Sorry, should have said - also did this with the spares. I can see no >>>>>> drives on 'head1' sysconfig -r that are 'wrong' and likewise on 'head2'. >>>>>> >>>>>> >>>>>> Checked this with grepping 'partner' and ensuring they're all the >>>>>> right pair of controllers. (On both sides) >>>>>> >>>>>> >>>>>> >>>>>> On 11 June 2015 at 11:56, Borzenkov, Andrei < >>>>>> andrei.borzenkov(a)ts.fujitsu.com> wrote: >>>>>> >>>>>> If you just performed “replace” without un-/re-assigning drives that >>>>>> became spare, this is expected. Spares still belong to controller and will >>>>>> block auto assignment. >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> *From:* toasters-bounces(a)teaparty.net [mailto: >>>>>> toasters-bounces(a)teaparty.net] *On Behalf Of *Edward Rolison >>>>>> *Sent:* Thursday, June 11, 2015 1:09 PM >>>>>> *To:* toasters(a)teaparty.net >>>>>> *Subject:* Disk auto assign >>>>>> >>>>>> >>>>>> >>>>>> I've just added some new shelves to a (7-mode) filer pair, bring them >>>>>> up to two stacks of 10 shelves. >>>>>> >>>>>> I'm trying to get 'autoassign' working, and it isn't. >>>>>> >>>>>> I have: >>>>>> - swapped drives such that each 'head' has only disks from one >>>>>> 'stack'. >>>>>> >>>>>> Head 1 has 0a/5b >>>>>> >>>>>> Head 2 has 5a/5d >>>>>> >>>>>> >>>>>> >>>>>> I've had to manually trigger a small number of 'replace' operations. >>>>>> And then switched on autoassign in the options. >>>>>> >>>>>> I _think_ that should be it, but ... there's something I'm missing, >>>>>> because autoassign isn't working. >>>>>> >>>>>> >>>>>> >>>>>> Does anyone have suggestions for why this might be, or what else I >>>>>> could check? (I'm wondering if there's a misassigned but 'failed' drive >>>>>> that otherwise isn't showing up?) >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>> >>>>> >>>>> _______________________________________________ >>>>> Toasters mailing list >>>>> Toasters(a)teaparty.net >>>>> http://www.teaparty.net/mailman/listinfo/toasters >>>>> >>>>> >>>> >>> >> >

1 0

Multiple (more than 2) controllers in a cluster
by Philip Gardner, Jr. 11 Jun '15

11 Jun '15

Hi all - We are currently building out a new space (long time in the making), and some employees have come and go in between the process. We ended up with a 2240, a 3250, and an 8040. The 8040 is fully updated with CDOT 8.3. We are now preparing to do the same to the 2240 and 3250. The 2240 has the 10Gb mezzanine card, and the 3250 has two 10Gb PCI cards. I'm still learning, basically as I go, but does it make any sense to put all these filers in the same cluster? If not, does it make sense to use the CN1610's for each of the filers' cluster connections? I'm trying to think of a reason the 2240 needs to be in the same cluster, but can't really think of one. The 3250...maybe, since it is a decent amount of power. I know from some of the docs I've seen that the 2240 can use one of the 10Gb ports to connect to the other head in a non-redundant fashion, while still allowing the use of the other 10Gb ports for the data connection. Googling for things about multiple filers in the same cluster is difficult, or no one is talking about it. Thoughts or ideas? -- GPG keyID: 0xFECC890C Phil Gardner

3 4

Max number of hosts.
by Brad Thompson 11 Jun '15

11 Jun '15

Hi all, Has anyone ever seen these messages or know what causes them: lmgr.hosts.maxedout: Number of hosts active 50000 has reached the maximum number of hosts supported 50000. No additional host structures can be created. Thanks! Brad

3 3

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

toasters