Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
Gonna Send It
Jul 8, 2010

movax posted:

Yeah — I saw the name and was like "... but where are the optics??". Guess they didn't learn anything from ThunderBolt's foray into that. Good heads up on the connector fragility — don't think these see a lot of cycles but if I do end up with a board with those connectors, most seem to be vertically oriented so maybe I can add some epoxy or similar to help out.

Got any links to recommended cables / adapters that are good from a SI POV?

You also might want to make sure you can get the cables, last I checked Supermicro was out of stock and Asrock didn't even sell them directly.

Adbot
ADBOT LOVES YOU

movax
Aug 30, 2008

Gonna Send It posted:

You also might want to make sure you can get the cables, last I checked Supermicro was out of stock and Asrock didn't even sell them directly.

Ugh, this is going get me down a rabbit hole of looking up the PCB mount / other part numbers and building some PCBs...

mobby_6kl
Aug 9, 2009

by Fluffdaddy
I've been transferring data from old drives that were just sitting on the shelf to my newish NAS, and one of them isn't cooperating. It's a WD Caviar WD5000AAKS. It spins up and then stops before it's detected, maybe 10-15 seconds or so. I've seen bad sectors or drives just not spinning up, but not this. Is there a typical cause for this type of failure? I'm using an external USB adapter from a shucked drive so kind of hoping it'll work in a PC but preparing for the worst

priznat
Jul 7, 2009

Let's get drunk and kiss each other all night.

movax posted:

Yeah — I saw the name and was like "... but where are the optics??". Guess they didn't learn anything from ThunderBolt's foray into that. Good heads up on the connector fragility — don't think these see a lot of cycles but if I do end up with a board with those connectors, most seem to be vertically oriented so maybe I can add some epoxy or similar to help out.

Got any links to recommended cables / adapters that are good from a SI POV?

We usually buy cables from one of a couple vendors, serialcables.com and CS electronics.

They have a lot of funky options like splitting up the drives so you can have 2 x2 drives or the host connectors for dual ported drives.

HalloKitty
Sep 30, 2005

Adjust the bass and let the Alpine blast

priznat posted:

Also the “optical” part of oculink is a terrible lie, there are no optical cables for it at all. :mad:

OK, that was boring, I hadn't heard of it before, and I'd hoped for an optical+copper connection à la thunderbolt's original design.
The name therefore makes no sense and disappoints me.

priznat
Jul 7, 2009

Let's get drunk and kiss each other all night.

HalloKitty posted:

OK, that was boring, I hadn't heard of it before, and I'd hoped for an optical+copper connection à la thunderbolt's original design.
The name therefore makes no sense and disappoints me.

Tell me about it! That is one thing we’ve been searching for, someone to make a oculink -> optical fibre module but nada. Ended up making our own pcie edge connector to samtec module version. Oculink even has the power wires for a module just no one seems to be interested in making one.

It’s basically a glorified displayport connector.

BlankSystemDaemon
Mar 13, 2009



movax posted:

I fully plan on doing TrueNAS / BSD-based ZFS, so no worries there — I've never used ZFS on anything besides Solaris (lol) and FreeBSD via FreeNAS.

I'm familiar with U.2, but I've seen mention of U.3 which basically makes it, if I understand correctly, one connector, one type of cabling (I guess they picked a nominal impedance and stuck with it) and the controllers / PHYs on either end have to pass onto logic that can handle SAS, SATA or NVMe. Seems like it's rolling out with some of the newer PCIe 4.0 based NVMe drives like the Kioxias I mentioned earlier.
If it was me, I would probably stick with the Illumos-derived ZFS on FreeBSD 12 for now, and keep my data off OpenZFS.
I guess it's pretty ironic, since I was one of the first among many to adopt ZFS on FreeBSD back in the 7.0 days - but now I'm not moving onto OpenZFS until I'm forced to by 12.x going EOL.

That seems like an absolutely bonkers setup, and I can't wait to find out how many ways the state machine involved in training that can fail. :allears:

movax
Aug 30, 2008

BlankSystemDaemon posted:

If it was me, I would probably stick with the Illumos-derived ZFS on FreeBSD 12 for now, and keep my data off OpenZFS.
I guess it's pretty ironic, since I was one of the first among many to adopt ZFS on FreeBSD back in the 7.0 days - but now I'm not moving onto OpenZFS until I'm forced to by 12.x going EOL.

That seems like an absolutely bonkers setup, and I can't wait to find out how many ways the state machine involved in training that can fail. :allears:

Oh I might be misunderstanding then — is FreeNAS / TrueNAS OpenZFS, and not the Illumos-derived one? My last machine was OpenSolaris I think (openIndiana?), I just knew that Linux implementations of ZFS were all kind of 'meh' but thought that BSD / Solaris ones were relatively safe and simple.

And yeah — I was thinking about how many ways the PCIe LTSSM could break and upon further reading, they "cheat" (slash, weren't dumb) and just use GPIOs / straps to ID which protocol to speak. Glad someone won the argument on using pins on a connector for that, because goddamn the flow chart for auto-detecting ordered sets / training sets to figure out if you're SATA, SAS or NVMe and the right clock speed would be awful.

GreenBuckanneer
Sep 15, 2007

How are the synology NAS devices with running a VM?

Seems you can "unofficially" slap a 16gb stick in the models that say you can't, which would make running a VM worthwhile.

Being a plex server and also being able to run vms from it, while having it backed up to the cloud is getting me a bit excited about it.....though it'll be months before I can save up the money. Kind of debating on the DS920+ tbh

UCS Hellmaker
Mar 29, 2008
Toilet Rascal
Loving the fractal 804, was super easy to get everything together and it seems to tolerate my absolutely horrible cable management well. Things only meant to storage and backup so I tossed an I3 in it because I was more concerned with usable then VM usage. Right now its two shucked seagate 12tbs and two toshiba gold 8tbs with plans to expand as I need space and add redundancy. Running truenas core because I liked the web interface better then unraid personally. so far besides a learning curve its been pretty nice and already have the drives mapped to my computer. Tonight will be getting it in a permanent home and getting stuff transferred over. Then setting up the plex server on it so I can have everything working off of the Nas upstairs.

Nitrousoxide
May 30, 2011

do not buy a oneplus phone



In my time playing around with TrueNAS I found that ZFS was way more straightforward than ext4. There's no messing around with different share folders or having to create discrete pre-sized partitions for something that you might have to change down the line. You just have a pool and create your data sets in that pool and they grow or shrink as necessary. You can add or remove data sets as you need. It's easy to back up a whole pool at once rather than having to do multiple backups of each data set or each shared directory.

I really like the snapshot feature as well as it's a really space efficient way to provide redundancy against accidental deletions. Although obviously it's not a backup of any sort.

It's probably the thing I miss the most now that I'm running open media vault.

Actuarial Fables
Jul 29, 2014

Taco Defender

UCS Hellmaker posted:

Loving the fractal 804, was super easy to get everything together and it seems to tolerate my absolutely horrible cable management well.

+1 for the 804. There's a few issues I have with it like having to pull four drives out to get to the one you actually want to service, but for a "I don't really know what this system is going to end up doing" kind of build it's great.

BlankSystemDaemon
Mar 13, 2009



movax posted:

Oh I might be misunderstanding then — is FreeNAS / TrueNAS OpenZFS, and not the Illumos-derived one? My last machine was OpenSolaris I think (openIndiana?), I just knew that Linux implementations of ZFS were all kind of 'meh' but thought that BSD / Solaris ones were relatively safe and simple.

And yeah — I was thinking about how many ways the PCIe LTSSM could break and upon further reading, they "cheat" (slash, weren't dumb) and just use GPIOs / straps to ID which protocol to speak. Glad someone won the argument on using pins on a connector for that, because goddamn the flow chart for auto-detecting ordered sets / training sets to figure out if you're SATA, SAS or NVMe and the right clock speed would be awful.
FreeNAS 11 uses the Illumos-derived ZFS implementation that's in FreeBSD 7 through 12. TrueNAS 12 and FreeBSD 13 (not yet released, only available if you build through sources or use a development snapshot) uses the LLNL-derived OpenZFS implementation which used to be called ZFSonLinux and has become the standard implementation worked on by the OpenZFS project.
The Illumos-derived implementation in FreeBSD will continue to be worked on until FreeBSD 12 goes EOL some time in June, 2024 if everything goes to plan.

Incidentally, NetBSD implements the Illumos-derived implementation too, by way of the port they made of FreeBSDs implementation - so it's possible they'll eventually also move to the OpenZFS standard implementation.

ZFS on Mac and ZFS on Windows, as well as Illumos-derived OS', are similarly expected to move to the OpenZFS standard implementation, at which point ZFS will likely be the most portable modern filesystem, ever.

Going with pins over anything else makes it seem like they actually know what they're doing. Makes me wonder how much arm-twisting it took.

GreenBuckanneer posted:

How are the synology NAS devices with running a VM?

Seems you can "unofficially" slap a 16gb stick in the models that say you can't, which would make running a VM worthwhile.

Being a plex server and also being able to run vms from it, while having it backed up to the cloud is getting me a bit excited about it.....though it'll be months before I can save up the money. Kind of debating on the DS920+ tbh
Motherboard manufacturers only test a certain number of DIMMs to find out what the highest supported memory amount is, but if you're lucky and/or pay through the nose, you can often get more memory by having DIMMs which fit in the tolerances, allowing you to go beyond what the manufacturer says.
DIMMs with chips from SK Hynix and Micron are the ones I've got the best experience with, when it comes to getting more memory on a board than is officially supported - but it's by no means a science, and typically comes down to sourcing an exact model and revision of memory.

Brain Issues
Dec 16, 2004

lol

GreenBuckanneer posted:

How are the synology NAS devices with running a VM?

I can’t speak for the newest models, but I’ve got the DS1618+, ram upgraded to 16gb, and a 1TB SSD cache. VM performance with Windows 10 and Ubuntu is awful. Do not bother. The CPUs in the SOHO/Personal Synologys just aren’t powerful enough to run VMs smoothly.

Things I use my Synology for:
Storing Data
Downloading content (Sonarr, Download Station)
VPN Server
Time Machine Backups
Run a Wordpress website

Things I don’t use my Synology for:
VMs
Plex Server (Transcoding performance is too poor)

Brain Issues fucked around with this message at 14:17 on Jan 14, 2021

Axe-man
Apr 16, 2005

The product of hundreds of hours of scientific investigation and research.

The perfect meatball.
Clapping Larry
DS1618+ doesn't have hardware transcoding so you are basing it totally on the cpu which honestly is not going to have the power that plex would like for it.

If you are going vms for windows, a rack xs+ unit for a xeon decent processor is going to what you need for it.

In production we have an RS3617xs+ and a cluster with an RS3613xs+ in HA with ram expansion and ssd cache and run a EXSI server for about 26 windows vms.

Brain Issues
Dec 16, 2004

lol

Axe-man posted:

DS1618+ doesn't have hardware transcoding so you are basing it totally on the cpu which honestly is not going to have the power that plex would like for it.

If you are going vms for windows, a rack xs+ unit for a xeon decent processor is going to what you need for it.

In production we have an RS3617xs+ and a cluster with an RS3613xs+ in HA with ram expansion and ssd cache and run a EXSI server for about 26 windows vms.

Yeah, I know. That's why I specified these issues only apply to the SOHO/Personal units.

GreenBuckanneer was contemplating buying the DS920+ and talking about using it as a Plex Server and for VMs. Neither of which I recommend doing.

GreenBuckanneer
Sep 15, 2007

Brain Issues posted:

Things I use my Synology for:
VPN Server

I hear it's good for that, and also as an email server. The latter piqued my interest, as it would be nice to self-own an email server and email address, if that's possible.

Also maybe also hosting a wordpress website, though I don't know if you'd have to also still pay for wordpress 🤔

Nitrousoxide
May 30, 2011

do not buy a oneplus phone



Axe-man posted:

DS1618+ doesn't have hardware transcoding so you are basing it totally on the cpu which honestly is not going to have the power that plex would like for it.

If you are going vms for windows, a rack xs+ unit for a xeon decent processor is going to what you need for it.

In production we have an RS3617xs+ and a cluster with an RS3613xs+ in HA with ram expansion and ssd cache and run a EXSI server for about 26 windows vms.

I'm personally considering whether to get a little 9U rack that I can double up as a table since they're about 20" tall, or a pretty typical table height, for next to my desk. Then I can relocate my networking gear into there, and build a 4U rack mounted NAS/Server to handle my storage/Plex/Home Hosting needs. I guess it would depend on how quiet I can make a 4U mounted computer.

Then I can put my existing mATX desktop on top of the rack (which is itself just on a side table at the moment).

Nitrousoxide fucked around with this message at 15:23 on Jan 14, 2021

EC
Jul 10, 2001

The Legend

Axe-man posted:

DS1618+ doesn't have hardware transcoding so you are basing it totally on the cpu which honestly is not going to have the power that plex would like for it.

What would be a good Synology pick for transcoding then? Most of my stuff will be direct play but I want to be sure I can transcode stuff for the future.

Axe-man
Apr 16, 2005

The product of hundreds of hours of scientific investigation and research.

The perfect meatball.
Clapping Larry
Ironically, the transcoding NAS are all the lower end residential, but, i would recommend for best performance, the following units from personal experience:DS720+ , DS920+ or DS1520+ all have same processor, but the DS1520+ has 4 ethernet ports for link aggregation if you need multiple users and 8 gigs of ram instead of 4 in the DS920+ and 2 in the DS720+. They have hardware transcoding, and are using Gemini Lake refresh.

If you have only a few users you can get away with a DS218+, DS418+ , in my opinion though expect less performance than the others.

The other option is to just handbrake to convert all your videos into a format that anyone can use.

Axe-man fucked around with this message at 16:46 on Jan 14, 2021

Teabag Dome Scandal
Mar 19, 2002


If I have an xfs array in Unraid there isn't really a progressive way to upgrade drive capacity without replacing all of them at once, correct? The other file system is the one that does mismatched drive sizes and even formatting a new drive as that file system won't change the overall array type?

HalloKitty
Sep 30, 2005

Adjust the bass and let the Alpine blast

Teabag Dome Scandal posted:

If I have an xfs array in Unraid there isn't really a progressive way to upgrade drive capacity without replacing all of them at once, correct? The other file system is the one that does mismatched drive sizes and even formatting a new drive as that file system won't change the overall array type?

I'm pretty sure unraid only allows you to use its 'dedicated parity disk(s)' array, at least I've never seen anything else. Whether the drives themselves are formatted individually as XFS or BTRFS is not important. Upgrade bit by bit at will. The only thing you need to be aware of is that the parity disk(s) need to be larger than or as large as the largest data disk. The documentation is pretty OK, and I'm sure there are helpful people on the unraid forums if you need any further in-depth help

HalloKitty fucked around with this message at 21:11 on Jan 14, 2021

Teabag Dome Scandal
Mar 19, 2002


HalloKitty posted:

I'm pretty sure unraid only allows you to use its 'dedicated parity disk(s)' array, at least I've never seen anything else. Whether the drives themselves are formatted individually as XFS or BTRFS is not important. Upgrade bit by bit at will. The only thing you need to be aware of is that the parity disk(s) need to be larger than or as large as the largest data disk. The documentation is pretty OK, and I'm sure there are helpful people on the unraid forums if you need any further in-depth help

yup, I'm familiar with all of that (the plan was to upgrade the parity drive first) but the Unraid documentation specifies mismatched drive size use was a BTRFS feature. I know the array should probably support the drive but I'm assuming with an xfs array the extra space won't be utilized unless all the drive sizes match?

Or maybe I'm misunderstanding what the documentation is saying? Right now I have 3 drives and 1 parity drive all at 3tb each (9tb usable space). I have space for one more drive and 5tb seems about the sweet spot right now and that was going to be the new parity drive while using the old parity drive to expand the array (4 drives, 1 parity, 12tb usable space). My concern is that later I won't be able to swap out a 3tb drive for another 5tb drive and use that extra 2tb (14tb total) without the other 3 drives also being 5tb.

HalloKitty
Sep 30, 2005

Adjust the bass and let the Alpine blast

Teabag Dome Scandal posted:

I'm assuming with an xfs array the extra space won't be utilized unless all the drive sizes match?

Nah, in fact, XFS is (was?) the default in unraid, so that's really the most tested config. I'm in a unique position to confirm the behaviour, as I actually built two identical (and I mean identical, same silverstone case, supermicro board, RAM, hard drives, PSU) unraid NASes, and I formatted one with BTRFS, and one with XFS. I've also replaced drives several times (to increase capacity). They behave identically, with regards to parity and mix and matching drive capacities. You get all the space either way

HalloKitty fucked around with this message at 22:15 on Jan 14, 2021

CopperHound
Feb 14, 2012

Teabag Dome Scandal posted:

Or maybe I'm misunderstanding what the documentation is saying?
Pretty sure you are over thinking it. Just toss the drive in there and you can use it all minus the difference between your parity and next largest drive.

Teabag Dome Scandal
Mar 19, 2002


HalloKitty posted:

Nah, in fact, XFS is (was?) the default in unraid, so that's really the most tested config. I'm in a unique position to confirm the behaviour, as I actually built two identical (and I mean identical, same silverstone case, supermicro board, RAM, hard drives, PSU) unraid NASes, and I formatted one with BTRFS, and one with XFS. I've also replaced drives several times (to increase capacity). They behave identically, with regards to parity and mix and matching drive capacities. You get all the space either way

CopperHound posted:

Pretty sure you are over thinking it. Just toss the drive in there and you can use it all minus the difference between your parity and next largest drive.

Rad. Thanks everybody!

One more quick question. I'm maxed out on SATA ports so I need to buy an expansion card. I'm using a Lenovo TS140. I only have space for 1 more drive so it doesn't need to support many SATA ports. Whats a good stable card to grab?

Teabag Dome Scandal fucked around with this message at 22:34 on Jan 14, 2021

priznat
Jul 7, 2009

Let's get drunk and kiss each other all night.

Teabag Dome Scandal posted:

Rad. Thanks everybody!

One more quick question. I'm maxed out on SATA ports so I need to buy an expansion card. I'm using a Lenovo TS140. I only have space for 1 more drive so it doesn't need to support many SATA ports. Whats a good stable card to grab?

Cheapest/easiest would be check amazon or wherever for "sata expansion card", looks like 2-4 ports is in the realm of about $30USD. They'd all be using the marvell cheapy controllers which would be fine.

bit more spendy/expandable option would be check amazon/ebay for LSI/broadcom cards like the SAS9211-8I HBA which gives you 8 SATA ports internally. I think there are OEM versions that are even cheaper that require a FW flash to go to HBA mode, iirc that was discussed here before so someone who got one can add on.

Either option would require a free PCIe slot, x1 for the first option and x8 for the 2nd.

Rooted Vegetable
Jun 1, 2002

Teabag Dome Scandal posted:

If I have an xfs array in Unraid there isn't really a progressive way to upgrade drive capacity without replacing all of them at once, correct? The other file system is the one that does mismatched drive sizes and even formatting a new drive as that file system won't change the overall array type?

There are several answers that could apply - I'm summarising a few ways of many to solve this.

Firstly, make sure you have an independent backup of any important data in case anything goes wrong.

If you have a parity drive and if your new drive(s) is larger than any other drive, and you have physical space, do a parity drive swap and use the old parity drive as a normal storage drive.

If you have run out of physical space and want to swap a drive you have a few options. Again if you have parity (and it's up to date), you could simply take out the old drive, put in the new one, and do a rebuild of the drive onto the new, presumably larger, drive.

For the next two methods, you'll do well get the unBalance plugin (for large unraid friendly file movements) and a file manager like Krusader (useful for checking files ended up in the right place)

If you don't have parity, no physical space free but do have some free space in your existing array, you could:
1. Either globally (global share settings) or for each and every share (each share's settings page) - exclude the disk you will remove. (discussion about that)
2. Use the unBalance plugin to scatter all the data on the disk to be removed, which moves it all to other drives.
3. Once the disk to be removed is empty, physically remove it and swap in the new disk.
4. Undo the exclusion from step 2 so that it can be used again (don't forget this bit).
5. Boot up and ensure the new disk added to the array in your "Main" page in the dashboard.
6. Optional: If you have a share you particularly want on that exact disk, use unBalance again to gather files onto it.

If you don't have parity, no physical space free and no free space in your existing array:
1. Connect your new disk by USB (for instance, before you shuck it) and add it to your array in a spare slot.
2. Either globally (global share settings) or for each and every share (each share's settings page) - exclude the disk you will remove. (discussion about that). Make sure the new disk is included.
3. Use the unBalance plugin to scatter all the data on the disk to be removed, which moves it all to other drives (including but not limited to your new one).
ALTERNATIVE to steps 2-3: use Krusader to move the files directly from the disk to be removed to the new disk that will remain.
4. Once the disk to be removed is empty, physically remove it and swap in the new disk.
5. Boot up and ensure the new disk remains in the array in your "Main" page in the dashboard. You may wish to remove the vacant slot from the now absent disk.
6. Undo the exclusion from step 2 so that it can be used again (don't forget this bit)

Teabag Dome Scandal
Mar 19, 2002


priznat posted:

Cheapest/easiest would be check amazon or wherever for "sata expansion card", looks like 2-4 ports is in the realm of about $30USD. They'd all be using the marvell cheapy controllers which would be fine.

bit more spendy/expandable option would be check amazon/ebay for LSI/broadcom cards like the SAS9211-8I HBA which gives you 8 SATA ports internally. I think there are OEM versions that are even cheaper that require a FW flash to go to HBA mode, iirc that was discussed here before so someone who got one can add on.

Either option would require a free PCIe slot, x1 for the first option and x8 for the 2nd.

Nothing to avoid? That is mainly my concern since I've seen people talk about unstable cards and I don't want to accidentally get one that is incompatible or the lovely chipset of the year. As long as it isn't going to gently caress something up I don't need the fanciest thing.

Rooted Vegetable posted:

For the next two methods, you'll do well get the unBalance plugin (for large unraid friendly file movements) and a file manager like Krusader (useful for checking files ended up in the right place)

ugh I swear I looked for a plugin like unBalance when I initially set this thing up and can't recall finding anything so thanks for the link!

It also sounds like I was over complicating things so I likely won't need your helpful instructions until I actually replace a data drive but I really can just take a new drive, copy poo poo from the old drive onto it, and slot it in and everybody is going to be happy? Or am I only doing that if I don't have parity? Is using a parity rebuild the recommended way to repopulate an upgraded drive?

priznat
Jul 7, 2009

Let's get drunk and kiss each other all night.

Teabag Dome Scandal posted:

Nothing to avoid? That is mainly my concern since I've seen people talk about unstable cards and I don't want to accidentally get one that is incompatible or the lovely chipset of the year. As long as it isn't going to gently caress something up I don't need the fanciest thing.

Something like Startech would probably be fairly decent, there are some chinese brands on amazon claiming to use the same marvell chip as well. 9 times out of 10 it probably comes from the same factory.

The LSI HBA would be the spendiest but it's a well known quality brand and you have the option of adding more drives in the future if you want as well.

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness

Teabag Dome Scandal posted:

Nothing to avoid? That is mainly my concern since I've seen people talk about unstable cards and I don't want to accidentally get one that is incompatible or the lovely chipset of the year. As long as it isn't going to gently caress something up I don't need the fanciest thing.

The Marvell ones generally won't gently caress your data up, they're just slow. That said, LSI-based cards are Professional grade vs the cheapest poo poo Marvell could put out, sooooo, yeah. They're also not much more expensive--you can find them on eBay any day of the week for <$50. Just search for LSI 9211. Like this one which for $50 gets you a card that should "just work" and a pair of break-out cables to plug your drives into.

The upside of going with a HBA like that is you can carry it along to a new build and hang a lot of drives off it, meaning lots of expandability in the future.

Also worth noting that, while HBAs like that technically use a 8x slot, if you've got a server/workstation motherboard that has the backs of the PCIe slots cut out (or you're adventurous and have a Dremel), you can shove it into a 4x or even a 1x slot as long as you don't need the extra bandwidth (even a 1x slot is more than enough for a single drive).

priznat
Jul 7, 2009

Let's get drunk and kiss each other all night.
Yeah and also don't worry that the 9211s are pcie gen2, it's totally fine and will work with any pcie slot as long as it can fit the x8 (x16, x8 or shorter with open ended back like DrDork mentioned)

I would also recommend the 9211 or other decent HBAs over the startechs but just wanted to mention the cheapest/simplest option :)

Teabag Dome Scandal
Mar 19, 2002


priznat posted:

Something like Startech would probably be fairly decent, there are some chinese brands on amazon claiming to use the same marvell chip as well. 9 times out of 10 it probably comes from the same factory.

The LSI HBA would be the spendiest but it's a well known quality brand and you have the option of adding more drives in the future if you want as well.

DrDork posted:

The Marvell ones generally won't gently caress your data up, they're just slow. That said, LSI-based cards are Professional grade vs the cheapest poo poo Marvell could put out, sooooo, yeah. They're also not much more expensive--you can find them on eBay any day of the week for <$50. Just search for LSI 9211. Like this one which for $50 gets you a card that should "just work" and a pair of break-out cables to plug your drives into.

The upside of going with a HBA like that is you can carry it along to a new build and hang a lot of drives off it, meaning lots of expandability in the future.

Also worth noting that, while HBAs like that technically use a 8x slot, if you've got a server/workstation motherboard that has the backs of the PCIe slots cut out (or you're adventurous and have a Dremel), you can shove it into a 4x or even a 1x slot as long as you don't need the extra bandwidth (even a 1x slot is more than enough for a single drive).

I don't really mind paying a little extra so I might just go that route and grab one off ebay. Thanks for the help!

Rooted Vegetable
Jun 1, 2002

Teabag Dome Scandal posted:

It also sounds like I was over complicating things so I likely won't need your helpful instructions until I actually replace a data drive but I really can just take a new drive, copy poo poo from the old drive onto it, and slot it in and everybody is going to be happy? Or am I only doing that if I don't have parity? Is using a parity rebuild the recommended way to repopulate an upgraded drive?


You could do almost any of them if you meet the pre-requisites for each process. Yes, if you have the room and space you could add the drive to the array first, copy the files to it using your preferred method, then remove the old drive from the array then physically remove it. Note I used a variation on the specifics of your words, but same spirit.

To the other part of the question, you can do that if you do or do not have parity, it's really up to you. I just gave you a few options. The Unraid forums (any of the links above) likely have more options and proceedures too if you're looking.

I won't come down on if it's recommended to use the rebuild or not, I don't personally have a parity drive, but I have seen it at least been raised as a valid option.

BlankSystemDaemon
Mar 13, 2009



priznat posted:

The LSI HBA would be the spendiest but it's a well known quality brand and you have the option of adding more drives in the future if you want as well.
$27 is "spendiest"? Because that's what I found a M1015 for on eBay with free international shipping.
EDIT: And here is another M1015 for $24, also with free international shipping.

I guess the first is slightly more expensive if it includes the bracket, but if that's what breaks the bank, then I don't know what to say.
They are fantastically cheap for what you get.

BlankSystemDaemon fucked around with this message at 03:20 on Jan 15, 2021

movax
Aug 30, 2008

BlankSystemDaemon posted:

$27 is "spendiest"? Because that's what I found a M1015 for on eBay with free international shipping.
EDIT: And here is another M1015 for $24, also with free international shipping.

I guess the first is slightly more expensive if it includes the bracket, but if that's what breaks the bank, then I don't know what to say.
They are fantastically cheap for what you get.

That's a SAS2008 right? Probably the lowest-end / cheapest HBA I'd throw at FreeNAS and it's easily appears to have the most run time / validation time thrown against it. The fact that they didn't HW-lock / prevent reverting to IT mode is a godsend; they could have easily done something with eFUSEs / something... Marvell-esque like that.

I almost said Broadcom-esque, but the SAS3xxx series I think were all developed after the Broadcom acqusition.

BlankSystemDaemon
Mar 13, 2009



movax posted:

That's a SAS2008 right? Probably the lowest-end / cheapest HBA I'd throw at FreeNAS and it's easily appears to have the most run time / validation time thrown against it. The fact that they didn't HW-lock / prevent reverting to IT mode is a godsend; they could have easily done something with eFUSEs / something... Marvell-esque like that.

I almost said Broadcom-esque, but the SAS3xxx series I think were all developed after the Broadcom acqusition.
Yep, it's the old LSI SAS2008 controller - IBM used to ship ServeRAID M1015 as a value-add for every server sold.

Also, here's a M1015 with SFF-8087-to-SATA breakout cable for $42.

EDIT: And here's a list of all the SAS2008 based controllers so you don't have to rely on it being IBM branded if you don't care about that (you really shouldn't, it makes no difference if you're not putting it into a HP or Dell server, if memory serves.

priznat
Jul 7, 2009

Let's get drunk and kiss each other all night.
On ebay the LSIs are still more than the cheapo marvell sata expansions which can be had for $10 :haw:

But yeah a HBA for $30 is a no brainer if the seller looks reasonably reputable.

EC
Jul 10, 2001

The Legend

Axe-man posted:

Ironically, the transcoding NAS are all the lower end residential, but, i would recommend for best performance, the following units from personal experience:DS720+ , DS920+ or DS1520+ all have same processor, but the DS1520+ has 4 ethernet ports for link aggregation if you need multiple users and 8 gigs of ram instead of 4 in the DS920+ and 2 in the DS720+. They have hardware transcoding, and are using Gemini Lake refresh.

If you have only a few users you can get away with a DS218+, DS418+ , in my opinion though expect less performance than the others.

The other option is to just handbrake to convert all your videos into a format that anyone can use.

It's a shame the DS1520+ only has 5 bays. I was looking at the Ds1821+ but it seems to have the same basic specs as the DS1621+ just with two more drive bays.

Adbot
ADBOT LOVES YOU

H110Hawk
Dec 28, 2006

EC posted:

It's a shame the DS1520+ only has 5 bays. I was looking at the Ds1821+ but it seems to have the same basic specs as the DS1621+ just with two more drive bays.

If you're getting into the territory where you have more users transcoding than space and network you shouldn't run plex on the NAS. Ironically the ram is unlikely to have much impact on plex workloads unless your plex ram usage itself for the database or whatever balloons up. Using it as block cache on large media files isn't going to save you unless you have small files (1gb/hr) and very high hot spotting (lots of users watching the same few things.)

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply