Sun Storage 7110

[Duke]

Hi Chris, I set my shares up on the SAN as everyone full control them re-set them all in Windows by doing \\SAN then right click on the share and set them manually.

 

I have on my student areas a group for staff and gave them special permisisons for read and execute everything from that folder down inc sub folders and files and it seemed to work fine on my testing, staff don't know they have this access, I put it in place so it was there for if needed was my idea rather than having to re-tweak them. I would screenshot but awaiting my box to come back up from the Q3.1.0 software which has the fix in for my crashing problem

 

Haven't noticed the share permissions on mine but will check in the morning for you....

 

Cheers mate, much appreciated! I forgot I could set security settings on the top-level share through Windows as long as it's not a hidden share. Looks like that'll probably be my best option currently.

 

If you could have a look at the share permissions that'd be really handy. Maybe it's just me but it was a bit odd.

 

EDIT: Looks like the doc apaton pointed to agrees with you:

 

Therefore, ACLs should be managed at either the filesystem level or the share level, but not both. To simplify management of the Sun Storage 7000 Unified Storage System in a pure Windows environment, it is recommended to set both the share-level permissions and filesystem-level permissions to full control for everyone, and then manage the individual permissions from a Windows client.

 

Chris

Edited October 2, 2009 by Duke

[Duke]

Interesting thread comparing NetApp and Sun storage here on the Ars forums: NetApp or Sun Storage

 

A lot of people playing with serious hardware frequent those forums so it's nice to see Sun being generally recommended over the NetApp solutions. Time for NetApp to reconsider their licensing prices...?

 

Chris

[john]

Thought I would post here, the new software seems to have cured the bug I had with the last release, no crashes all day Yay! Monday will be a 2nd day of testing, although the full backups of all data on it over the weekend should give it a good test as well so we will see what happens on Monday before counting all the chickens.

 

Did you get your permissions sorted Chris from the guide book or do you want a screenshot still?

[Duke]

Thought I would post here, the new software seems to have cured the bug I had with the last release, no crashes all day Yay! Monday will be a 2nd day of testing, although the full backups of all data on it over the weekend should give it a good test as well so we will see what happens on Monday before counting all the chickens.

 

Did you get your permissions sorted Chris from the guide book or do you want a screenshot still?

 

Thanks for the update, sounds good to me (except the BE issue you noted elsewhere)! Assuming nothing comes up then I'll do the upgrade on Thursday or Friday I think. I haven't got around to doing the permissions completely yet but I think I'll go through what the Sun doc recommends so I know I'm following the best practices. Since they say to do root OR share I'll probably leave the reset share permissions as-is, double-check my root permissions, then use a Windows box to set up the rest.

 

Cheers,

Chris

[pete]

Is anyone else responding to the latest bout of HP salespeople with "Sun will give me two servers for the price of one and a half-price san, what can you offer me?"

 

None of them have stepped up to match the offer yet.

[Duke]

Is anyone else responding to the latest bout of HP salespeople with "Sun will give me two servers for the price of one and a half-price san, what can you offer me?"

 

None of them have stepped up to match the offer yet.

 

Haha, good times... I'm not in a buying position until next summer, but I'm going to guess Dell won't be able to match that. Even without the BOGOF the Dell R710 with 48GB RAM came out as more expensive than the equivalent Sun box.

[john]

Not had that call, but that would be interesting as I'm hoping next financial year to replace 3 servers that are all hobbling around the place, so any offers for those would be welcome

[pete]

I tried the shadow filesystem migration for the first time on the 7110 this morning (create share, specify which NFS mount to migrate it from during share setup) and it fails with an error of: "Data migration source: Failed to mount filesystem"

 

The NFS mountpoint is available read-only (as per the migration doc) and mounts fine on another system. It's on an OpenFiler 2.3 box at servername:/mnt/folder1/folder2/folder3 and my mount line in the BUI is servername:/mnt/folder1/folder2/folder3

 

I've tried swapping hostname for IP without any success. There aren't any restrictions on the openfiler box that would prevent the 7110 mounting the share (bunch of .isos and random junk used to test how quickly migration works). Looking in the logs on the Openfiler, the 7110 attempts to do a lookup and never attempts to mount the share:

 

Attempted mount from 7110

Oct  9 09:33:31 openFiler nmbd[3500]: [2009/10/09 09:33:31,  0] nmbd/nmbd_namequery.c:query_name_response(108) 
Oct  9 09:33:31 openfiler nmbd[3500]:   query_name_response: Multiple (2) responses received for a query on subnet OpenfilerServerIP for name DOMAINNAME<1d>. 
Oct  9 09:33:31 openfiler nmbd[3500]:   This response was from IP 7110-IP, reporting an IP address of DOMAINCONTROLLERIP.

 

Compare this with a successful mount from a linux server

Oct  9 09:06:57 openfiler mountd[14275]: authenticated mount request from linuxserverIP for /mnt/folder1/folder2/folder3 (/mnt/folder1/folder2/folder3) 

 

What am I missing here? Given the Openfiler box, linux server and 7110 pull dns from the same server I'm slightly baffled as to why it can't mount the share. I haven't tried the CLI yet.

 

Ok, CLI gives me:

littlerascal:shares XenStorage/openfiler-test (uncommitted)> commit
error: invalid property value
      "nfs://servername/mnt/folder1/folder2/folder3" for "shadow":
      Failed to mount filesystem
error: failed to commit changes

 

Using the syntax example show in TFM.

 

And I'm on the latest firmware, which fixes the .zfs wierdness.

Edited October 9, 2009 by pete

[apaton]

Does openfiler allow root mounts ?

 

The SS7000 probably mount the system as root. NFS root user has the lowest privileges unless specified.

 

 

Example NFS options are usually root= or anon=0

 

 

Andy

Edited October 9, 2009 by apaton

[apaton]

For people not moving to Q3 patch yet, Sun have released 2009.Q2.5.1

 

Sun Storage 7000 Series Software Updates - Fishworks - wikis.sun.com

 

Andy

[pete]

ZFS Deduplication : Jeff Bonwick's Blog

 

*pokes Phil*

[Duke]

ZFS Deduplication : Jeff Bonwick's Blog

 

Whoop whoop, de-dupe!

[pete]

Yup, I have a stack of vm images with maybe 1-2Gb of differences each, so eagerly awaiting this. Plus availability of enterprise rated 2TB drives + ZFS dedupe makes me much happier about rebuilding the backup server.

[Soulfish]

Looking forward to dedpue

[Duke]

Has anyone got any docs / whitepapers / blueprints on the recommended network configuration of the S7000 in different environments? I've just got a single 1Gb feed going to my 7410 right now and it's handling the current load fine, but I need to start planning for heavy CIFS use, medium NFS VM use plus maintaining performance of the web interface.

 

I'm thinking:

 

• 2x 1Gb trunk on a public (i.e. same as rest of network) IP address for clients directly accessing CIFS shares
  
• 1x 1Gb feed on a private IP range that ESXi boxes will talk to via VMkernel for NFS access to VMs
  
• 1x 1Gb for access to the BUI should CIFS usage get really heavy or the trunk fail
  

 

How do people normally handle this with regards to hostnames and DNS? Would I give the IP of the trunk feed the same name in DNS as my current single feed has, then make a new DNS entry called something like 'sanmgmt' for the single feed I want to use the BUI over, and just put that 'sanmgmt' name into my browser?

 

The VMware/NFS feed will be on a private IP range so should I just specify this and access to the ESXi boxes via IP address rather than hostname?

 

EDIT: Also, anyone got any suggestions/advice on handling redundant cabling, so Port 1 on Device A goes to Switch 1, Port 2 on Device A goes to Switch 2, Switch 1 and Switch 2 are interconnected, and Switch 1 and Switch 2 are both connected to separate ports on Device B, so that there's no single point of failure? Kind of like this, although I'm sure you know what I mean. How do you normally handle the IP addressing and hostnames for something like that?

 

Cheers!

Chris

Edited November 3, 2009 by Duke

[apaton]

REF - Cabling - You have two technologies to choose from IPMP and LACP.

 

LACP 802.ad will effectively load balance over interfaces

IPMP will provide failover (with a small performance increase on outbound traffic)

 

For best Availability & Performance

You will need a redundant switch stack (eg Cisco 3750/Nortel 5500) and just use LACP.

 

If you have two independent switches

You could just use IPMP, but you could use both LACP and IPMP together. See example here

Problem is that you will run out of NIC ports quite quickly.

 

Andy

[john]

On my 7110 I have 3x1GB LACP links for CIFS (I only use it for CIFS at presnet) and then the last 1 GB Link is to our backup LAN where our NAS boxes and backup stuff lives so I can actually backup in the day as well as night without causing delays or too much load (not that it wouldn't manage to cope as it flys)

[Duke]

Cheers, I'd planned on using LACP for the links but will investigate IPMP too. On that diagram you linked to, which IP address would actually be used to communicate with the SAN? Does IPMP tie the two LACP links together and the LACP connections themselves aren't actually used directly?

 

John or Andy, how do you then handle DNS hostnames (or IP addresses) for the separate links? Obviously the SAN itself has to have a hostname, do you bind a different DNS name to the 'other' IP address that you're using for something else (e.g. backup in your case John) or just specify the other link via IP address on the backup server?

 

Many thanks, it's really appreciated.

Chris

[cookie_monster]

We have 3xNICs in an LACP link and 1 that is currently used for management but could be used for CIFS. The LACP link is carrying all of our iSCSI traffic for our VMs, 2x File servers, 2x DC's, Citrix and a couple of test boxes and it hardly breaks a sweat.

[apaton]

Cheers, I'd planned on using LACP for the links but will investigate IPMP too. On that diagram you linked to, which IP address would actually be used to communicate with the SAN? Does IPMP tie the two LACP links together and the LACP connections themselves aren't actually used directly?

 

You would use the IPMP address, which can fail-over between aggregates.

 

DNS/IP addresses, I'm flexible on this, just depends what's accessing the devices.

 

• NFS/CIFS clients I would create a DNS entry for the IPMP device ( if no IPMP then aggregate.)
  
• VMWARE (iSCSI/NFS) and replication between devices, then I'm happy with IP addresses.
  

 

Andy

[john]

Cheers, I'd planned on using LACP for the links but will investigate IPMP too. On that diagram you linked to, which IP address would actually be used to communicate with the SAN? Does IPMP tie the two LACP links together and the LACP connections themselves aren't actually used directly?

 

John or Andy, how do you then handle DNS hostnames (or IP addresses) for the separate links? Obviously the SAN itself has to have a hostname, do you bind a different DNS name to the 'other' IP address that you're using for something else (e.g. backup in your case John) or just specify the other link via IP address on the backup server?

 

Many thanks, it's really appreciated.

Chris

The 3x trunk has just 1 ip and DNS Name so its SAN-01 in our case (original I know!)

The 1x standalone one again has 1 IP (different range completly) and I just access it via IP IE \\172.16.10.200\share_name and it works fine. It does appear on that LAN as SAN-01 though so its not a problem at all for us.

[Duke]

Thanks all, very much appreciated!

 

I'll use IP addresses if necessary, but if I did go for multiple DNS hostnames (e.g. sancifs, sanmgmt) does anyone know if this would cause communication problems with the box as the hostname being used to communicate is different from the hostname specified on the device? I have some software/hardware that always refers to the 'local' hostname and will redirect you to that when you click a link on the BUI even if the address you originally went to was different.

 

Cheers,

Chris

[Butuz]

Hey Guys.

 

What mirroring / redundancy options are available on a pair of 7110 2tb SANs?

 

I originally bought a pair of these SAN's based on a project that was to create a new building with a second server room and have a SAN in server room building 1 and a san in server room building 2 and both of them mirroring - but the entire new building plan got shelved, and so I am left with both SAN's in the same rack of my only server room. *sigh*

 

Not ideal!!! But better than nothing.

 

This storage is only being used for VM's over NFS. (no CIFS or iSCSI). So what do you knowledgable guys thing is the best way for me to configure it so that my VMs are mirrored across both SANS so that if one san does go down I will be able to access the VM's off the second SAN straight away. I see in the interface there is "Replication" and also on the filesystem there is "Snapshots" but I don't know what does what. This is what I am aiming for. Can it be done?

 

I have not had time to do proper research into getting this configured on 7110, I have 6 big projects on at the moment as well as the day to day stuff, and Offstead in in January. As you can imagine I am struggling somewhat!!!! If anyone can help me my giving me the lowdown or idiots guide on this - I will be most greateful.

 

Cheers!

 

Butuz

[Ric_]

@Butuz: Replication is what you want. You can do scheduled replication and (IIRC) asynchronous replication which is instant-ish.

[Duke]

Did you buy through Cutter and are you on their forums? I only ask because I just asked an almost identical question on there a couple of days ago!

 

Replication will copy the data between two SANs on a schedule of your choosing. I'm not sure how well this will handle data on running VMs.

 

Replication will transfer the data and metadata in a project and its component shares either at discrete, point in time snapshots or continuously. Discrete replication can be initiated manually or occur on a schedule of your own creation. With continuous replication, data is streamed asynchronously to the remote appliance as it's modified locally at the granularity of storage transactions to ensure data consistency. In both cases, data transmitted between appliances is encrypted using SSL. A project replicated on another appliance is an exact copy of that local project. Every share, share property, snapshot, and configuration setting is replicated.

 

Snapshots do just as the name suggest and will 'snapshot' the data giving you an instant full backup. Changed data is written to new blocks, non-changed data is referenced from your snapshot so that a snapshot backup doesn't take up and room until the data has been changed.

 

A snapshot is a point-in-time copy of a filesystem or LUN. Snapshots can be created manually or by setting up an automatic schedule. Snapshots initially consume no additional space, but as the active share changes, previously unreferenced blocks will be kept as part of the last snapshot. Over time, the last snapshot will take up additional space, with a maximum equivalent to the size of the filesystem at the time the snapshot was taken.

 

I don't think either of these technologies will give you the instant failover you want, for that you need clustering which unfortunately isn't supported on the 7110. Replication will give you failover ability but I think the actual 'activation' of the backup machine would need to be done manually.

 

The Sun Storage 7000 series Unified Storage System supports cooperative clustering of appliances. This strategy can be part of an integrated approach to availability enhancement that may also include client-side load balancing, proper site planning, proactive and reactive maintenance and repair, and the single-appliance hardware redundancy built into all Sun Storage 7000 series appliances. Because the clustering feature relies on shared access to storage resources, it is available only on the Sun Storage 7310 and 7410. You will be unable to configure clustering on other appliance models, or if the two heads are not of the same model.

 

I plan to cluster two 7410's across the site next summer to do what you're talking about.

 

Chris