Disaster strikes as NAS3 crashes

This past weekend we had a power brownout for about 4 hours. This caused my servers to fail-over to battery power. The batteries don’t last long with servers running. I guess something went sour with the automatic shutdown of my NAS3 which is used only for my VMware virtual machines and it did an improper shutdown. The RAID has crashed.

I don’t have anyone to blame other than myself and I knew eventually this day would come. NAS3 was in RAID-0. That means striping with no redundancy. A failed array on RAID-0 typically means total data loss. I take daily backups of this entire NAS nightly so I am aware and prepared for the risk of using striping. That does not mean that it’s a fun time recovering from it.

Adding additional redundancy for blackouts

Currently, one of the hardest things to recover from in my current home-lab environment is a total power blackout. Everything right now is planned & designed around losing certain components like 1 disk, 1 switch/network cable, etc. However when everything is off and I need to bring things back online it’s a painstaking and very manual process. Over time my environment has also become more and more complex. This latest outage has me scratching my head at how to recover faster & simpler from a power blackout.

Continue reading…

Whoops! My Samsung 950 PRO M.2 512GB SSD experience

Samsung-950-Pro-SS

This story is a bit of a sad story. Mainly due to the failure of my own personal oversight over a very important specification requirement and snowballed by my online shopping addiction.

For a little back story – I had been reading great reviews all over the internet about Samsung’s new 950 PRO 512GB M.2 NVMe SSD drive having amazing benchmark speeds. This drive features cutting-edge V-NAND-based NVMe SSD supports PCI Express Gen 3 x4 lanes, providing a higher bandwidth and lower latency to process a massive amount more data than SATA SSDs. It outperforms SATA SSDs by over 4.5 times in sequential read and by over 2.5 times in sequential write, delivering the speeds of 2,500 MB/s read and 1,500 MB/s write respectively. I have never purchased or used a M.2 drive before, so this would be me venturing into a bit of an unknown territory.

I wanted to get my hands on one of these really badly! I knew my motherboard (Gigabyte X99 G1) had a M.2 slot so I checked the website to confirm and at a glance everything appeared to be right… So I pulled the trigger and purchased it. Patiently waited 2 weeks for it to arrive. Finally, I was able to install the drive after some frustration with BIOS settings and drivers. But something wasn’t right… I was concerned as to why I was getting a little under half of the speed other people were getting on their benchmarks. What gives?

Well… lets take a closer look at my mother board specifications again. It is the Gigabyte X99 G1:

000250_2016-01-22 09_17

See that..!? Notice the “PCIe x2/x1“? The Samsung 950 PRO 512GB is a PCIe 3.0 x4. So that means that on my motherboard the M.2 drive is operating at literally under half the bandwidth that it’s capable of running at. Only at this point, having the drive purchased and installed, did I realize that there is a major difference between “M.2”  and “Ultra M.2” which also known as PCIe 3.0 x4 which supports up to 32 Gb/s (4 GB/s) compared to PCIe 2.0 x2 which would only support 8 Gb/s (1 GB/s).

nooooooooo

That means for a M.2 drive using PCI-E 2.0 x2, it only has a 25% potential speed increase over SATA III. Not exactly what I was going for considering the $329 USD price I paid. This means for me to be able to use the drive at full speed I will need to find a motherboard with a LGA2011-v3 CPU socket and a Ultra M.2 socket. A harsh lesson to learn and a fix I will have to put off for some time. Luckily I may be able to just get a M.2 x4 to PCIe adapter to fix this problem.

In the end I can only laugh at myself as a technical professional to make a mistake like this – but hey this is my story and hopefully it may help you avoid making the same mistake. Check to ensure your motherboard’s M.2 interface version supports PCIe 3.0 x4  NVMe 1.1!

 

Update 1/28/2016:

Fixed! Thanks to the $30(CAD)  Addonics ADM2PX4 M2 PCIe SSD to PCIe 3.0 4 Lane Adapter.

2016-01-28 16_48_52

The Samsung Magician software is also showing that I am using all four lanes, instead of just two!

2016-01-28 17_02_32

Here’s a bonus benchmark of another Samsung drive I have with RAPID mode (RAM cache) enabled. :O
2016-01-28 16_42_33

 

Do you have a story to tell about your M.2 drive experences? Let me know in the comments below!

Storage Refresh 2016 – Time to Build! (Part 2)

IMG_0254

Part 1: http://www.vskilled.com/2015/07/storage-refresh-2016-the-plan/

The hard drives have arrived today from NCIX and it’s now time to build it out to finally increase the storage capacity in my home lab. I’ve made only minor changes to the original plan; I ended up shying away from the Seagate 8TB archive hard drives I had originally planned on buying to use strictly for backup purposes. Much like 3TB drives, I just don’t have any confidence in them long-term.

DeviceCurrent Drive LayoutCurrent CapacityDesired Drive LayoutDesired Capacity
NAS 1 (Thecus N5550)5 x 2TB (RAID 5)7.21TB5 x 4TB (RAID 5)16TB~
NAS 2 (Thecus N5550)1x4TB, 2x2TB, 1x1TB, 1x640GB (JBOD)8.9TB5 x 2TB (RAID 5)7.25TB~
VMH01 (Dell C1100)1 x 1TB1TB2 x 8TB, 2 x 2TB (JBOD)20TB~

The end result stays the same. I’m looking to end up with two large data/media storage pools with about 22TB of usable storage. A considerably large increase from my existing data capacity of 7.21TB.

The challenge now is performing a safe and successful data migration to the new storage. Normally I use NAS2 as my backup/archive NAS. I am going to remove the drives from it and move some of them into my VMH01 (Dell C1100) and create a temporary datastore to backup of all the data on there. That way I can safely create a backup of the data and still be semi-protected by RAID.

After some careful scavenging through some documentation I found that I would probably be able to swap the disks from NAS2 and move them into NAS1 without losing any configuration or data. However this is risky, so I will store a 3rd copy of my data on a JBOD  on “BackupSrv”. In this case the risk is worth the reward if it pays off because I will be saving myself from having to copy the data from the BackupSrv JBOD again, and worst case scenario I still have the data on the drives so I can just swap then back if I needed to roll-back the change.

The Step-by-Step Action Plan:

  1. NAS2: Destroy JBOD, power-off, remove drives
  2. VMH01: Add 1x4TB, 1x2TB, 1x1TB. (Add to BackupSrv VM)
  3. BackupSrv: Create JBOD datastore for backup
  4. NAS2: Add 5x4TB NAS drives, build raid, re-configure NFS, rsync, ftp, etc
  5. NAS1: Full rsync backup to BackupSrv & NAS2, verify data
  6. Power-off both NAS1 and NAS2.
  7. Swap disks from NAS2 into NAS1, NAS1 into NAS2. Power on, cross fingers.
  8. Verify data and shares. It works!
  9. Data Migration Completed
  10. Cleanup: Reconfigure Rsync Backup Schedules
  11. Cleanup: Update Home Lab page, CMDB, Wiki
  12. Cleanup: Permissions on shares

* – Veeam Backup Repository moved temporarily to NAS1 (approx 600GB~)
* – NFS datastores + permissions will be lost during a RAID rebuild
* – Printer Scan-to-FTP Setup

 

Lets take a look at our storage now:

000204_2015-11-09 08_47

 

Excellent. Now I have a large RAID5 14.5TB share for media/data storage, another RAID5 7.26TB share for more data storage, and another 7TB of disks in JBOD for archive/backups. I have a LSI MegaRaid MR SAS 9260-8i 8 Port SAS Raid Card on the way to properly archive/backup JBOD the drives so that I can present the disks more cleanly to a backup VM.

Storage Refresh 2016 – The Plan

homelab-bottom

The time has come to increase storage capacity in the home lab. I expect that before the end of this year that I will have less than 1TB of free space left on my primary data NAS. That is a problem, and an expensive one at that. At the time of this writing I have 1.66TB of 7.21TB free (77% used). My data growth rate is currently between 3-5% on average per month. That gives me about 2-3 months before I’m in a critical state.

Adding storage to the primary data/media pools means also means adding storage to the backup pools. You won’t catch me without a backup – you only need to be burned by that once before you learn that harsh lesson. Seagate has come out with a 8TB drive meant for backups only which will help with backup capacity. Overall have been pretty skeptical of these 8TB drives. It is strongly advised not to use them in a RAID setup, they use SMR (shingled magnetic recording) that allows the tracks on the platter to be layered on top of each other to increase platter density or tracks per inch (TPI). With that said they seem to be fairly robust. While one could argue that I could (should?) delete some stuff, I strongly disagree. I am a data hoarder. Do you literally throw out all your books after you’re done reading them? Probably not. Same goes with data.

Upgrading the primary NAS means I’ll need to rebuild RAID arrays, use NAS 2 as “swing” storage, move data onto the upgraded NAS 1, rebuild NAS 2, and so on. This will take a couple of days of just moving data around and ensuring I have a backup at all times. During the swing process I am particularly vulnerable to drive failure. Currently my backup NAS 2 is in a JBOD configuration. If any one of the drives fail during this read/write intensive transfer process – game over. For that reason I will be making a second backup onto the 8TB seagate drive, just in case.

The plan is to switch NAS 1 into 5 x 4TB RAID 5, NAS 2 into what NAS 1 is currently (5 x 2TB RAID 5). I’ll then be leveraging my VMH01 (Dell C1100) for the backup pool drives (2 x 8TB, 2 x 2TB in JBOD) served up by a NAS4Free virtual machine. To help wrap my head around what I am doing I like to draw things out on my whiteboard. Here is my “draft” design. Apologies for the chicken scratch.

storagerefresh2016-draft

I’ll be re-purposing an existing 4TB drive in NAS 2 and moving it into the NAS 1 raid pool (hence why only purchasing 4 x 4TB drives as seen below). This saves me the cost of buying another 4TB drive.

I will be using a multi-vendor setup using a mix of Seagate and Western Digital drives. That will make things a little more robust in the long term. Currently I just have desktop rated drives in the primary NAS which, by manufacture guidelines, are only rated for a maximum of 2 in RAID 1/0 and they are also only rated for 8×5 use. The WD Red and Seagate NAS series drives are designed for use in home NAS and servers. They offer a good price to performance ratio, and possess a few features which make them more suitable for RAID arrays such as TLER, higher vibration tolerance (which should result in a longer lifespan), consume less power and are rated for 24/7 use.

western-digital-red-4tb

DriveQuantityCost (CAD)
$1,459.94
Seagate ST8000AS0002 8TB 5900RPM 128MB Cache SATA3 Archive Hard Drive OEM - for Backup Data Only2 x $319.99 ea$319.99
Western Digital WD WD40EFRX 4TB Red SATA3 6GB/S Cache 64MB 3.5in Hard Drive2 x $209.99 ea$419.98
Seagate ST4000VN000 4TB 64MB SATA 6GB/S 3.5IN Internal NAS Hard Drive2 x $199.99 ea$399.98

All said and done I will end up with two large data/media storage pools with 22TB~ of usable combined storage.

A considerably large increase from my existing data capacity of only 7.21TB. The idea being for this to last at least 3+ years. NAS 2 which is currently a JBOD for backups only will now be another usable RAID5 protected data pool. Each NAS backed up to VMH01’s backup storage JBOD.

DeviceCurrent Drive LayoutCurrent CapacityDesired Drive LayoutDesired Capacity
NAS 1 (Thecus N5550)5 x 2TB (RAID 5)7.21TB5 x 4TB (RAID 5)16TB~
NAS 2 (Thecus N5550)1x4TB, 2x2TB, 1x1TB, 1x640GB (JBOD)8.9TB5 x 2TB (RAID 5)7.25TB~
VMH01 (Dell C1100)1 x 1TB1TB2 x 8TB, 2 x 2TB (JBOD)20TB~

I’ll be sure to post updates with pictures on the build and upgrade process when the time comes. For now I’ll be trying to saving up some cash to make this plan come together.