Will this work? X11SCL-F with Intel Core i3-8300 FreeNAS build

s00hr7

Dabbler
Joined
Feb 12, 2021
Messages
11
Hey folks, I've been reading through this forum for a while and came up with a build for a FreeNAS system, I'd like you to take a quick look at.

Hardware:
In my country, there is almost no option to get older/used hardware, so I have to stick with buying new stuff.

Intended Usage:

There is nothing that needs transcoding, the majority will be Photos, Music, and some other non-media files. The space required is not that much at the moment, so I want to start with just two mirrored disks with the option to add another mirrored pair.

Besides the storage, there will be other applications running in jails:
Mailserver
PhotoPrism (which can get quite CPU intensive when a lot of images are run through the ML categorizers, but that's not time-critical)
Some file hosting service. I am not a fan of Nextcloud and I also don't need all the features, so I am looking for something less bloated.
Eventually, some VMs just to play around with some stuff.

Specific Questions:
  • The Supermicro case comes with a 250W PSU, would that be enough even with 4 disks?
  • Please double-check if CPU/Memory/Board work together and all support ECC (after going through all the threads about matching setups I got a bit paranoid about that)
  • Any obvious errors in the setup that I've missed?

Thanks a lot!
 

ChrisRJ

Wizard
Joined
Oct 23, 2020
Messages
1,904
Regarding the matching of board, CPU, and RAM: I am not sure what you expect. Either the components are specified by the respective vendors to work together or not. Supermicro generally provides that information, so that should answer your question. Or do you mean something different?
 

s00hr7

Dabbler
Joined
Feb 12, 2021
Messages
11
Hi Chris, I was hoping to find someone who already tried this combination and can confirm that it works well together. The RAM I got from the Supermicro compatibility list, so that should work and the CPU should be supported and work with ECC as well.

I am just overly cautious here since, in this country, a shop can actually refuse to take back what you've ordered in case it doesn't work. And I really want to avoid sitting on a bunch of hardware that ended up not working together :)
 

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,175
Your chassis is too small, you'd need a miniITX board to use it. The X11SCL-F is microATX.

One other thing that sticks out is the 32 GB of RAM. It seems somewhat excessive for a pair or two of 4 TB disks. You do mention VMs, so I guess it's going to depend on your specific workload. Keep in mind that you can always start with 16 GB and add more later if it turns out that you do need more. You end up somewhat more limited at 64~112 GB of RAM, assuming you don't replace the 16 GB DIMM, but even 64 GB is plenty for most/all 4-disk scenarios.
 

LeDiable

Dabbler
Joined
May 6, 2020
Messages
36
I'm using that board with a Core i3-9100F. Currently, the 9100F is cheaper than the 8300, at least here in the US. If you don't need an integrated GPU, you could consider that CPU as well.
 

rvassar

Guru
Joined
May 2, 2018
Messages
971
One other thing that sticks out is the 32 GB of RAM.

Amusingly.... I was going to point out that the RAM might be too little for the future VM load...
 

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,175
Amusingly.... I was going to point out that the RAM might be too little for the future VM load...
16 GB go a long way with 1 GB VMs and not very far with 8 GB VMs.
 

rvassar

Guru
Joined
May 2, 2018
Messages
971
16 GB go a long way with 1 GB VMs and not very far with 8 GB VMs.

True... I had Zoneminder, Plex, Gogs, and a MySQL DB w/myphpadmin all stuffed into 16Gb at one point. It fit, and was stable, etc... But it displaced so much of the ZFS ARC that performance went in the dumpster. I really should upgrade to 32Gb, but my X9SCL motherboard is paying the "Mac Penalty" at this point in the curve. The 8Gb ECC UDIMM's are associated with the Mac Pro, and get the premium markup on the used market. I just can't bring myself to drop $140 on DDR3 RAM in 2021.
 

ChrisRJ

Wizard
Joined
Oct 23, 2020
Messages
1,904
I just can't bring myself to drop $140 on DDR3 RAM in 2021.
Just out of curiosity: For how much RAM would that be? I have a couple of X9SRi-F boards and found some 16 GB RDIMMs for 20 Euros a piece.
 

rvassar

Guru
Joined
May 2, 2018
Messages
971
Just out of curiosity: For how much RAM would that be? I have a couple of X9SRi-F boards and found some 16 GB RDIMMs for 20 Euros a piece.

The X9SCL doesn't used registered RAM, it uses UDIMM's. 4 x 8Gb goes for as much as $140 US even today...

I've actually toy'ed with the idea of finding another motherboard and moving to something that could use RDIMM's or LRDIMM's, but this hasn't been the year for it.
 

LeDiable

Dabbler
Joined
May 6, 2020
Messages
36
I perhaps foolishly bought 2x32GB sticks from Supermicro when I built my system. I was concerned that if I got 16 or 32 now, then decided to upgrade later on, the RAM would be harder to find & more expensive, so I just went to the i3's limit as part of the build. If in the future I upgrade to a Xeon I can increase RAM by just adding 2 more sticks, instead of having to remove whatever I originally bought. Probably bad logic based on my use case, but I'm going to run this for 5-10 years, so I'll consider it an investment.

And, I've got plenty of room for VMs I guess. A secondary Pi-Hole might be good, but I'd probably only give that 1-2 GB. Who knows what else might suit my needs.
 

s00hr7

Dabbler
Joined
Feb 12, 2021
Messages
11
Hi all, giving an update on the mentioned build. Hardware arrived a few days ago (I went with the Define R6 as a case instead of the Node 804 and also choose to use the 32G RAM from the start) and the assembly went fine. Installed TrueNAS Core and tried out TrueNAS Scale as well. Then decided to do proper Hardware validation as mentioned here.

I am not very familiar with the smart values, but these error rates for two brand new disks look awefully high (only ran a short and a long test on both disks):

Code:
SMART Attributes Data Structure revision number: 10
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   080   066   044    Pre-fail  Always       -       111632062
  3 Spin_Up_Time            0x0003   097   097   000    Pre-fail  Always       -       0
  4 Start_Stop_Count        0x0032   100   100   020    Old_age   Always       -       4
  5 Reallocated_Sector_Ct   0x0033   100   100   010    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000f   066   060   045    Pre-fail  Always       -       3684819
  9 Power_On_Hours          0x0032   100   100   000    Old_age   Always       -       28 (101 17 0)
10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail  Always       -       0
12 Power_Cycle_Count       0x0032   100   100   020    Old_age   Always       -       4
184 End-to-End_Error        0x0032   100   100   099    Old_age   Always       -       0
187 Reported_Uncorrect      0x0032   100   100   000    Old_age   Always       -       0
188 Command_Timeout         0x0032   100   100   000    Old_age   Always       -       0
189 High_Fly_Writes         0x003a   100   100   000    Old_age   Always       -       0
190 Airflow_Temperature_Cel 0x0022   069   066   040    Old_age   Always       -       31 (Min/Max 22/34)
191 G-Sense_Error_Rate      0x0032   100   100   000    Old_age   Always       -       0
192 Power-Off_Retract_Count 0x0032   100   100   000    Old_age   Always       -       0
193 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       19
194 Temperature_Celsius     0x0022   031   040   000    Old_age   Always       -       31 (0 22 0 0 0)
197 Current_Pending_Sector  0x0012   100   100   000    Old_age   Always       -       0
198 Offline_Uncorrectable   0x0010   100   100   000    Old_age   Offline      -       0
199 UDMA_CRC_Error_Count    0x003e   200   200   000    Old_age   Always       -       0
240 Head_Flying_Hours       0x0000   100   253   000    Old_age   Offline      -       28 (170 208 0)
241 Total_LBAs_Written      0x0000   100   253   000    Old_age   Offline      -       111270074
242 Total_LBAs_Read         0x0000   100   253   000    Old_age   Offline      -       361988



SMART Attributes Data Structure revision number: 10
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   080   066   044    Pre-fail  Always       -       111689762
  3 Spin_Up_Time            0x0003   097   097   000    Pre-fail  Always       -       0
  4 Start_Stop_Count        0x0032   100   100   020    Old_age   Always       -       4
  5 Reallocated_Sector_Ct   0x0033   100   100   010    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000f   065   060   045    Pre-fail  Always       -       3441997
  9 Power_On_Hours          0x0032   100   100   000    Old_age   Always       -       28 (215 70 0)
10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail  Always       -       0
12 Power_Cycle_Count       0x0032   100   100   020    Old_age   Always       -       4
184 End-to-End_Error        0x0032   100   100   099    Old_age   Always       -       0
187 Reported_Uncorrect      0x0032   100   100   000    Old_age   Always       -       0
188 Command_Timeout         0x0032   100   100   000    Old_age   Always       -       0
189 High_Fly_Writes         0x003a   100   100   000    Old_age   Always       -       0
190 Airflow_Temperature_Cel 0x0022   069   066   040    Old_age   Always       -       31 (Min/Max 22/34)
191 G-Sense_Error_Rate      0x0032   100   100   000    Old_age   Always       -       0
192 Power-Off_Retract_Count 0x0032   100   100   000    Old_age   Always       -       0
193 Load_Cycle_Count        0x0032   100   100   000    Old_age   Always       -       19
194 Temperature_Celsius     0x0022   031   040   000    Old_age   Always       -       31 (0 22 0 0 0)
197 Current_Pending_Sector  0x0012   100   100   000    Old_age   Always       -       0
198 Offline_Uncorrectable   0x0010   100   100   000    Old_age   Offline      -       0
199 UDMA_CRC_Error_Count    0x003e   200   200   000    Old_age   Always       -       0
240 Head_Flying_Hours       0x0000   100   253   000    Old_age   Offline      -       28 (20 159 0)
241 Total_LBAs_Written      0x0000   100   253   000    Old_age   Offline      -       102893394
242 Total_LBAs_Read         0x0000   100   253   000    Old_age   Offline      -       8796368


How likely is it that both of the disks are already faulty? Am I reading something wrong here?

The next thing I looked at was a CPU validation:

errors.jpg


As you can see, after about 5 minutes the CPU went up to 73°C (I am using the boxed cooler that came with the 8300) which I am not sure is a good value. It seems rather high for me. Could anybody share their experience if this is too high and I should probably use a different cooler?

The next thing was the fact that apparently there is no ECC available on the system. After a while, I realized that I ordered the wrong RAM (searched for the correct product ID, click on the first result, and ordered it, not realizing that the result didn't fully match, so I pretty much ordered the right RAM just without ECC, so this is on me... I'll get this replaced).

I finally stopped the test after it started complaining about smart errors from sdb (which is one of the two Crucial SSDs). So right now I got the feeling that this system SEEMS to run fine, but looking at it closer it seems to be a bunch of trash waiting to fail...

So TL;DR
  • Should I return the Seagate disks because of the high error rate?
  • Should I get a new CPU cooler for the 8300?
  • How can I investigate the smart errors on sdb? looking at "smartctl -A /dev/sdb" doesn't show any errors/failures
Update:
I ran mprime for about one hour, CPU temp was stable at around 80°C, running breakin now again and the temperature seems to be maxed out at 75°C. It is still reporting the smart errors for the SSDs. Since this is pretty old software (2015) is it possible that it just has problems with these SSDs and the messages are meaningless?
 
Last edited:

s00hr7

Dabbler
Joined
Feb 12, 2021
Messages
11
Update, just to add to the list of things that seem to be broken. Ran the memory test and got this:

memerror2.jpg


Am I cursed?
 

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,175
Since this is pretty old software (2015) is it possible that it just has problems with these SSDs and the messages are meaningless?
Very likely, yes, especially with SSDs.

Update, just to add to the list of things that seem to be broken. Ran the memory test and got this:

View attachment 45479

Am I cursed?
Memtest86+ is semi-abandoned and somewhat dubious for anything older than Sandy Bridge.
 

s00hr7

Dabbler
Joined
Feb 12, 2021
Messages
11
Update:

Regarding the Seagate Ironwolf disks, I found: https://www.truenas.com/community/t...st-raw_read_error_rate-seek_error_rate.68634/
In short: Seagate is using a proprietary data format for some of those SMART values. Using the commands mentioned in the thread, I get 0 errors, so the disks are fine (currently running badblocks on both of them, without any errors yet)

Regarding the SMART errors for the crucial SSDs there is this: https://www.truenas.com/community/t...te-194-temperature_celsius.77881/#post-542499
In short: The crucial firmware has an extremely low-temperature threshold (it triggered at about 45°C although the crucial documentation says they are fine up to 70°C, so this can be ignored. It's just a bit annoying since TrueNAS will alert me with a smart error, every time the SSDs get a little bit warmer.

I'll replace the memory with the ECC version and will try with a newer version of Memtest86. I searched through the forum, but there doesn't seem to be a conclusive recommendation on what memory testing software to use, since they all seem to not be very reliable.

It would still be interesting to know CPU temperatures for continuous 100% load from folks who have non-retail coolers to see if a different cooler could be an improvement here.
 

rvassar

Guru
Joined
May 2, 2018
Messages
971
It would still be interesting to know CPU temperatures for continuous 100% load from folks who have non-retail coolers to see if a different cooler could be an improvement here.

73/C is not out of spec, but I would attempt to address it. The Intel coolers seem to be hit or miss somewhere after Ivy Bridge. I get away with one, and never see temps above 55-60/C, but there seems to be a range of results. The only advice I can offer, use quality thermal paste in a uniform layer. If you consistently see temps above 60/C, start looking at other solutions.
 
Top