Start a Conversation

Solved!

Go to Solution

1397

March 20th, 2021 21:00

Live Volume - Between two SC7020

We are running two SC7020 and we are also utilizing Live Volume on most of our Volumes. Each SC7020 is also used with active VM's operating off each SC7020. Live Volumes xcross over both SC7020.. I have read that all writes will first be written too tier 1 storage, does this apply to the live Volumes. ? 

As for live volumes as they will only be used in the event of a failure with the Source SC7020, Can Live Volumes be moved to Tier 3 Storage ?. 

Also SNAPSHOTS are also active on the Secondary Live Volumes.. Is this necessary. ? 

Each nigh start at 7pm our SNAPSHOTS are process and the moving of data between the Tiers begin, this process takes more than 4.5-6 hours to complete and really loads up the SAN IOPS and Disk Activity.. I/O waits climbs to around 35-45.. All Active Servers are VM's. We also operate 24/7. Thoughts on how to improve.. 

Last question, we were not aware that SC7020 would use Dual Mirror, we were not aware this would eat up storage space by a 1gb uses 3gb storage, we have thinking to turn off Dual Mirror on our Tier 1 and Tier 2 storage, and or turn on compression.. Thoughts. 

 

 

 

 

4 Operator

 • 

2.3K Posts

March 26th, 2021 12:00

You can setup Single/Dual Redundancy on Tier level which means    Single Redundancy (RAID10/5) four your Tiers 1+2  and Dual (RAID6) only for your TIER3.

About compression..... the data reduction (compression or compression/dedup) is a post process so it never effects the writes and when it comes to reads its configurable because you can select if you wanna only compress the frozen blocks or all blocks. If you select frozen (which is the default) than only the outdated data within a snapshot and which is already changed is compressed/dedup.

We use dedup/compression for all data.

Regards,
Joerg

Moderator

 • 

7.6K Posts

March 22nd, 2021 09:00

Hello Ash61,

What is your current SCOS version? If you are using synchronous the data is written to tier1, if you are using asynchronous then you can select which tier you want the data to go to at the destination system.  Do you have deduplication enabled for live volumes?  Depending on your SCOS version you maybe able to change the bandwidth control for live volumes.  Here are the links to 2 different versions of the administrator guides as I am not sure of your SCOS version.  The guides list out the steps that is needed to make the changes. https://dell.to/3saQI65

 

https://dell.to/3c830Xu  

4 Operator

 • 

2.3K Posts

March 24th, 2021 00:00

LiveVolume sits on top of synchronous replication and for sure the secondary systems waits for sending the ACK until the data is saved.

The GUI doesnt prevent you to choose a different storage policy for your LV but we always use the same because in a event of a planned or unplanned downtime we would like to see same performance level for the Hosts(VMs). In theory you can sync between a AFA and a System with NLSAS.

We create autom. Snaps every view hours to trigger the "data progression on demand".

We never use DM on tier 1 and rarely on tier3*.  A rebuild with a SSD based tier 1 with the distributed spare feature enabled runs very quick and we think the possibility of a double disk failure is low. But when Mr. Murphy tries to kick in we have LV as plan B.

* We often have 10K SAS as tier3

Regards,
Joerg

6 Posts

March 26th, 2021 02:00

Additional, why does Dell SC Series use Double Mirror, it chews up so much additional space, and from my guest causes additional write I/O to th drives, not to mention RAID is being implemented by industry standard, than DM on Raid 5 and DM Raid will create additional parity  to be created

We intend to turn off DM and run normal Mirror and Raid 5,6 which the box seems to decide for itself to perform. 

I am finding it hard to find where all these saving come from in which the sale brochure speaks about.  told not to use compression as it will impact I/O even more.. And our tech does not recommend using dup and compression. We thought could be used... these apparently not if you want addition impact 

 

 

 

 

6 Posts

March 26th, 2021 02:00

Our Version is 7.4.21.4

Our Tier 1 is SSD, Tier 2 10k and Tier 3 7.2k.. Data Progression/processing of snap shots and the moving of data between tiers along with excessive RAID scubs which ramdonly appear at night and can run for 1 hour or 8 hours  kills any performance on our SC7020..  And it appear there is no way to stop RAID scubs.. I have never seen so many RAID scrubs performed on any storage below but our two sc7020 just about daily raid scubs are going on and on..  These processors really kill the SAN performance.. 

 

 

Moderator

 • 

7.6K Posts

March 26th, 2021 10:00

Hello Ash61,

With a compellent system you can set it up to be either single redundant or dual redundant.  When you use dual redundant then you will use dual mirror.  Here is the link to the administrator guide which explains a little more about the difference in redundantcy. https://dell.to/3d8GQDH

6 Posts

March 28th, 2021 00:00

Hi Sam, 

 

Can I turn off Snapshots on the volumes which are set as live volumes, it does make sense to me to be taking snap shoots on volumes which are Live Volumes. 

 

Please advise

 

 

Moderator

 • 

7.6K Posts

March 29th, 2021 10:00

Hello Ash61,

You can turn off snapshot on your other volumes if you would like too.  You will need to have snapshots enabled for using live volumes.

No Events found!

Top