14 Posts
0
1652
Expand storgage of Dell EqualLogic PS6100
Hi everyone, I want to expand my Dell EqualLogic PS6100. I am running it with raid 6, and 24 drives and each of them has 600GB and 1 hot spare. I just bought two 1.2TB drives, I want to ask that after I replace them 1 by 1, what will I need to do next? or can someone show me the whole process of this replacement? Thank you for helping and I really appreciate that.
dwilliam62
1 Rookie
1 Rookie
•
1.5K Posts
1
February 8th, 2023 10:00
Hello,
That is very old firmware. To be clear this is not a Dell supported method to upgrade drive capacity. You accept all risks to your data if there should be additional drives failed during so many rebuilds.
Re: Spare. Yes. that has to be the first one replaced. Otherwise it will simply rebuild on to the existing 600GB drive. When you replace the spare, then pull the first data drive, it will rebuild onto the 1.2TB drive but only use 1/2. That pattern will repeat itself until all the drives are replaced.
Regards,
Don
#iworkfordell
dwilliam62
1 Rookie
1 Rookie
•
1.5K Posts
0
February 9th, 2023 13:00
Hello,
Once you join them in the same pool, they will work together on the same volumes. The space will be balanced between the two members in relation to their size. Thie new one sounds like it will be much larger than the first one. Likely an "E" model with 7200RPM drives? Depending on the current load, and what you are using it for, now is a good time to decide do you split them or put them together in the same pool.
Re:20TB. The largest volume you can create on the PS SAN is 15TB. Depending on your OS you can either mount them using mount ponts, or use software striping to link the volumes at the OS level. Software striping does add complexity and typically can't be done while preserving existing data. Uness it was aleady in place. I.e. LVM, MD, Wndows stripping, etc... Using mount points in windows can be done and preserve data.
.i.e. current mount /data 10TB Volume
Add: /data/archive New 10TB volume or
/data/archive1 New 2TB volume
/data/archive2 Another new 2TB Volume
Great for storing year based data for example. You can add more over time
Depends on your needs and what OS you are using
re: Blank space. You can create additional volumes, use the new free space for snapshot reserve, etc.
The new volumes can be for your current servers or any new ones you add to it.
Back to deciding one pool with two member or two pools with one member each
If you do two in one pool, you will overall gain more IO performance. As you now have another controller, network ports and more drives and cache. I am guessing your current member is 10K RPM drives? If so, and if you have very latency sensitive needs, SQL, Excahnge, VDI, etc AND your current IO load is high, you might want to keep that member on its own. When you add the larger array with 7200 RPM drives, more of the volume data will reside on the larger member. IOs are split across them. IO operations won't complete until both arrays process them. The PS Series SAN, especially with current FW is designed to help offset this effect but you can't escape physics. More data will be on your slower member array.
If that 10TB volume is for archives, or other no latency senstive need then putting the two members in the same pool would be a good choice.
CON:
The only downside of the multimember pool option is the volumes are RAID-0 sriped across the members. If one array should completely fail, all the volumes will go offline until the failed member is restored. An uncoverable errror condition would mean data loss
Two pools with one member each:
In this configuration, you can move volumes from the original member to free up space and manually manage the IO balancing. I.e. most latency sensitive IO is on the 10K RPM member and the rest on the 7200RPM member in its own pool.
One benefit is a failure of one member does not impact the remaining member. Downside you have to manage voiumes yourself. It will take time to move volumes from one member to the other as it's done at a much lower priority.
I know that's a lot of info but I hope it helps you decide how you want to proceed.
It's very easy to go from two singe member pools to one muti-member pool. The "merge' process is quite fast. It will show as one in a few seconds. In the background space will be rebalanced between the members. New IO will automatically be spread across the members.
Regards,
Don
#iworkfordell
dwilliam62
1 Rookie
1 Rookie
•
1.5K Posts
0
February 7th, 2023 15:00
Hello,
Just replacing one or two drives will not increase the available storage. Data is stripe across all drives in equal mounts. Adding just two 1.2TB drives will force the array to only use 1/2 of those drives. It will use the lowest common denominator from all the drives available.
2nd. Where did you get those two drives? If they were not specifically coded for EQL arrays they will NOT work. THey can be the same part number, firmware and even Dell drives but without a speciic coding (which you can't add or bypass), the array will not use those drives.
Regards,
Don
#iworkfordell
briandohth
14 Posts
0
February 8th, 2023 06:00
thank you for the answer. I am going to buy a new 24 disks. I want to ask if after I replace them 1 by 1, will the system upgrade the storage and run on it by its self or do i need to take any action to do that? Thank you
dwilliam62
1 Rookie
1 Rookie
•
1.5K Posts
0
February 8th, 2023 07:00
Hello,
You are very welcome. Glad to be of assistance.
When you get those drives, test each one as the spare to make sure they are in fact EQL drives. Dell does not sell these drives anymore.
First, I would suggest backing up your data. You are going to be intentionally causing a faulted condition over and over again.
Replace the spare first then pull the first drive to be replaced. Make sure the rebuid completes.
Once that process is completed, the array will see the new capacity and begin the expansion process. The additoinal space will not be avaialble until the expansion completes. It will take some time as the process is run at a low priority so as not to interfere with onging IO.
What is the version of EQL firmware ?
I am also assuming this array is NOT under a Dell Support Contract? If so, this would void that contract.
Regards,
Don
#iworkfordell
briandohth
14 Posts
0
February 8th, 2023 09:00
i am currently using the version 7.0, I can confirm with you that all the drives are for the EqualLogic. By the way, is the first drive i need to replace is the spare one?
briandohth
14 Posts
0
February 8th, 2023 12:00
Thank you for the information, I just figured out another way to upgrade my storage, but I am not sure if it is going to work or not. Since the PS6100 is an extendable server, can I buy another PS6100 and link them together? will that work?
dwilliam62
1 Rookie
1 Rookie
•
1.5K Posts
0
February 8th, 2023 12:00
Hello,
That in fact is the method that is supported. It doesnt have to be another 6100. It just has to have enough capacity to hold the allocated space in the current 6100. It also has to be running a compatible version of FW. In your case 7.0 or 7.1. They also should be the same interface speed. GbE or 10GbE.
You would add it to the group in its own pool. Configure the network interfaces, and create the RAIDset and let the build complete.
There are two ways to migrate the data. 1.) Merge the members into the same pool 2.) Move volumes one by one.
My suggestion, is move ONE volume, that has data, to the new pool. This is to make sure the network is working properly. That tends to be the biggest issue.
Also if you have replication configured, let all replications complete then pause all inbound and outbound replication.
Assuming the test move goes, well you can either contine and move one volume at a time or do the merge.
If you merge pools, the process will show completed very fast. However, in the background data will begin being striped across the two members. I tend to let that balance complete. You will see the volume info will show volume space allocated across both members. If they are the same size, then 1/2 will be on each for example.
When you are ready you can either evacuate the original member into its own pool. This will leave all the data on the new member.
You can also leave them in this two member pool for even more space and greater IO capaciy as well. Since both members will process data for all the volumes by default.
If you do move it into its own pool you do have the option of deleting it from the group. Do this from the GUI, do NOT issue a reset on that member. That will leave the member orphaned in the group, it will always show offline.
The last option is if you want to keep the 1.2TB drives in the original chassis, you can shutdown the PS6100 Pull the 600GB drives. Put in the 1.2TB drives. Put the 600GB drives in the NEW chassis.
Making sure it's all cabled properly AND the firmware is the SAME on the new chassis. If it is different in anyway, including the minor version, the old array will NOT boot.
Regards,
Don
#iworkfordell
briandohth
14 Posts
0
February 9th, 2023 12:00
Thank you for your solution,
I want to ask you more about the blank space after I upgrade. Specifically, as I mentioned, I have 10TB on an existing volume, and I am about to buy another PS6100 with 96TB and link them together. According to my first goal, I just want to upgrade my existing volume to about 20TB, which is a part 10TB from the new 96 one. So the question is can I use 86TB left for another specific task? and how can i seperate the new PS6100 for 2 different use. Thank you
briandohth
14 Posts
0
February 10th, 2023 08:00
Thank you, William, this will be my last question regarding this upgrade. I know it's hard to explain here, but do you know how can I connect both of them on a physical aspect? The existing has 2 management ports and one of them is acting and has the green light flashing, should I plug the new switch with 2 management cables to the same switch i the existing connect to? Or what method i can do to link them 2 PS6100 together. Thank you so much
dwilliam62
1 Rookie
1 Rookie
•
1.5K Posts
0
February 10th, 2023 08:00
Hello,
You need to make sure that ALL the ports on BOTH members can connect to each other.
In a typical two physical switch configuration you alternate the EQL (data and management) ports to the two switches. Insuring that no controller or switch failure will result in all paths being down. So Member A Controller 0, port 0 goes to switch A. Member A, controller 0 port 1 goes to switch B. Member A, Controller 1, port 0 go to switch B, Member A Controller 1, port 1 goes to switch A
Same for the new member. You alternate the connections on both controllers AND the management ports. Mgmt port for CM0 goes to one switch, Mgmt port for CM1 goes to the other.
Both members need to be in the same VLAN and IP subnet. Otherwise you have to set up routing and without redundant routing a router failure could cause issues.
ALSO, the two switches MUST be connected together. If you are using VLANs for iSCSI make sure that the VLAN is set for that Interswitch Link (ISL). The ISL needs to be 70-80% of the total ports on the members. 6100's have four GbE ports, so at least 8x GbE ports or 2x 10GbE ports should be used to make up the ISL.
EQL presumes that any iSCSI port on any member can reach any other member and servers.
The Mangament port is the same way. So all those IPs should be in the same IP subnet and reachable by both members.
Please let me know if that explains it.
Regards,
Don
#iworkfordell
briandohth
14 Posts
0
February 16th, 2023 08:00
Thank you for the solution, but I will upgrade the existing drives. As you mentioned, I have to replace the spare one, and then the other drive. So, does that mean I need to replace two drives at a time? Or do I need to replace the spare first, wait for it to be rebuilt, and then replace the first drive? Finally, after the rebuild process is done, what function, or where can I click to officially expand the storage? Thank you
dwilliam62
1 Rookie
1 Rookie
•
1.5K Posts
0
February 16th, 2023 09:00
Hello,
ONLY pull ONE drive at a time. You remove the spare and replace it with the larger drive. You then either via the GUI or CLI make sure the drive came online as the new spare.
Then you can remove your first DATA drive. This will force a rebuild onto the new space you just installed. Then you replace that same slot that you pulled the data drive with the next NEW drive. That will become your new space. Again, verify that it came online as your spare.
When the rebuild finishes you can pull the next DATA drive and repeat the process. Add the new space back in and verify that it's rebuilding.
When all of the drives are installed, in the GUI or at the CLI, it will say "explanding"
When that completes the new space will be available.
Regards,
Don
#iworkfordell