XigmaNAS

I have seen some users(some newcomers) asking for some feature request lately with either related or non related to a "Network Attached Storage" Appliance, and some older ones that has been quite ignored and I will briefly explain here on why this requests are passed.

XigmaNAS Embedded ("Preferred Installation") runs entirely from a RAM Drive i.e a virtual disk drive allocated into the System RAM, this makes for the appliance to run from a temporary pseudo environment until a reboot/power loss occurs, so what this means ?, this meant that XigmaNAS Embedded uncompress all its applications, config files etc. etc. to the System RAM thus consuming a chunk from it, making it harder for the developers to maintain memory consumption efficiency while adding content, upgrading to major FreeBSD versions etc, not to mention the constant Embedded ROOT partition increase we need to do each time a bigger application/content is added, thus making the rest of the users to be forced to perform a new Re-install because the ROOT partition is too small error while trying to upgrade, and thats not good.

Let me note that this is not a design fault, but more exactly a known limitation inherited from the m0n0wall project(discontinued), and current devs are aware of, and no design changes/efficiency improvements will be made to current Embedded Platform any time soon.

To finalize, I will encourage users wanting extra applications to be included into XigmaNAS base to use jails, or just use OneButtonInstaller to easily add available Extensions withing the WebGUI such as Plex.

Alternatively advanced users can simply use a Full Platform which is meant for that specific purpose, adding/customize if you are already experienced.

To validate the Memory Consumption provided information, I will post a screenshots below showing the actual memory consumption side-by-side between Embedded and Full Platforms, based in 2GB system RAM and no services started, (note that the small disk space usage on the Full zroot platform is because of lz4 compression):



Regards

Always seemed a good thing to me. The original architecture of m0n0wall helped contain a minimal, manageable codebase, whilst supporting the new and more temporary features individual users want with minimal fuss. Also allows add-on developers to drop in and develop features they were motivated to build, without having to deep-dive into the project codebase as much (well, that's the ideal... )

Less complexity in the faces of those that don't want the add-on features
Less overhead in memory
Less code to maintain in the codebase
Less vulnerabilities in the codebase
Smaller attack surface
Encourages maintainers to look after add-ons and reduces project overheads
Etc.

And the downside is just, requiring an extra GB of RAM, and people having to add the custom features post-install.

Pu02 wrote: ↑

24 Oct 2018 06:47

Always seemed a good thing to me. The original architecture of m0n0wall helped contain a minimal, manageable codebase, whilst supporting the new and more temporary features individual users want with minimal fuss. Also allows add-on developers to drop in and develop features they were motivated to build, without having to deep-dive into the project codebase as much (well, that's the ideal... )

Less complexity in the faces of those that don't want the add-on features
Less overhead in memory
Less code to maintain in the codebase
Less vulnerabilities in the codebase
Smaller attack surface
Encourages maintainers to look after add-ons and reduces project overheads
Etc.

And the downside is just, requiring an extra GB of RAM, and people having to add the custom features post-install.

A bit late here, however just to add into, the memory consumption in Embedded Platforms will stay increasing regardless, while Full Platforms remains conservative, actually non an issue for most since memory is cheap nowadays, but may be an issue for Thin Clients and small devices with built-in limited memory.

Look at the updated screenshot above, RAM usage increased from ~39% to ~53% due latest feature additions plus system patches etc, (based in 2GB system RAM and no services started).

Regards

Hello Jose!

Had a question for you,

I do have 2 feature requests that I would love to see put into Xigma, but after reading your explanation, I have come to realize that this is not always easy to do.

I wonder, is is better for me to post what the suggestion on a new thread anyway, and see if you guys are able to integrate?

The 2 suggestions are:

1. HDSentinel or similar license free extension
2. Some sort of Benchmarking client/extension (Crystaldiskmark for example)

I can post to a new thread if you think its ok. What do you think?

sleyk2010 wrote: ↑

22 Nov 2018 03:43

Hello Jose!

Had a question for you,

I do have 2 feature requests that I would love to see put into Xigma, but after reading your explanation, I have come to realize that this is not always easy to do.

I wonder, is is better for me to post what the suggestion on a new thread anyway, and see if you guys are able to integrate?

The 2 suggestions are:

1. HDSentinel or similar license free extension
2. Some sort of Benchmarking client/extension (Crystaldiskmark for example)

I can post to a new thread if you think its ok. What do you think?

HDSentinel is a monitoring tool reading drive S.M.A.R.T. values. It is not available for platforms based on FreeBSD, source: How to: monitor Network Attached Storage (NAS) status

Crystaldiskmark is a windows-based benchmarking tool for disk drives and is not available for FreeBSD.

Well, S.M.A.R.T is already working in XigmaNAS. What more does HD Sentinel has to offer?
@sleyk2010 : Not judging your request, really asking because I have not used HD Sentinel.

P.S.: I find Diagnostics > Information > S.M.A.R.T. enough for my needs.
However now that I saw HD sentinel website, I am wondering (and have been for many years, just never proceeded with a feature request or something) if it would be easy for the devs to create a table in that page with a quick summary for all drives, above the detailed info
Perhaps something like (rough screenshot from excel)
HEALTHY or not can perhaps be determined by getting the info from SMART overall-health self-assessment test result or by checking absence of uncorrrectable/pending sectors or whatever
What do you think?

sleyk2010 wrote: ↑

22 Nov 2018 03:43

Hello Jose!

Had a question for you,
.....

Hi and welcome to the forum, what I have to say is pretty much the same what the previous post said.

P.S XigmaNAS already includes solid SMART monitoring/alerting features and you can even get emails reports about for convenience.

Regards

ChriZathens wrote: ↑

22 Nov 2018 14:30

Well, S.M.A.R.T is already working in XigmaNAS. What more does HD Sentinel has to offer?
@sleyk2010 : Not judging your request, really asking because I have not used HD Sentinel.

P.S.: I find Diagnostics > Information > S.M.A.R.T. enough for my needs.
However now that I saw HD sentinel website, I am wondering (and have been for many years, just never proceeded with a feature request or something) if it would be easy for the devs to create a table in that page with a quick summary for all drives, above the detailed info
Perhaps something like (rough screenshot from excel)

Capture.JPG

HEALTHY or not can perhaps be determined by getting the info from SMART overall-health self-assessment test result or by checking absence of uncorrrectable/pending sectors or whatever
What do you think?

Just to add, also Warnings/Attentions could be colored as ORANGE and not RED as it may overwhelm some users and they end just replacing a drive that just relocated a bare single sector, and that disk still could live forever, also is really annoying having a disk marked as RED just because a few harmless "UDMA_CRC_Error_Count" due a bad cable that has been replaced.

P.S As a fun fact, 80% of the really BAD hdd drives(assorted brands) I've replaced on customers Laptops/PC's, SMART have never shown any sight of damage, while they are clearly defective, either mechanically or electronically, and some with ridiculously low transfers speeds in the bytes ranges.

Regards

JoseMR wrote: ↑

22 Nov 2018 15:38

Just to add, also Warnings/Attentions could be colored as ORANGE and not RED as it may overwhelm some users and they end just replacing a drive that just relocated a bare single sector, and that disk still could live forever, also is really annoying having a disk marked as RED just because a few harmless "UDMA_CRC_Error_Count" due a bad cable that has been replaced.

P.S As a fun fact, 80% of the really BAD hdd drives(assorted brands) I've replaced on customers Laptops/PC's, SMART have never shown any sight of damage, while they are clearly defective, either mechanically or electronically, and some with ridiculously low transfers speeds in the bytes ranges.

Regards

Well, I totally agree with your points, however it always depends on what kind of error xigmanas reads from smart.. An example would be a disk with 1000 pending sectors which, while still in operating state, at least on a NAS should be scheduled for replacement quickly..
So I guess some stuff could be green, others orange and others red..

ChriZathens wrote: ↑

22 Nov 2018 15:53

Well, I totally agree with your points, however it always depends on what kind of error xigmanas reads from smart.. An example would be a disk with 1000 pending sectors which, while still in operating state, at least on a NAS should be scheduled for replacement quickly..
So I guess some stuff could be green, others orange and others red..

+1, Also agree with that, will be nice to just evaluate about what states/colors should be matched indeed, and leaving critical ones always RED.

Thanks Jose for the warm welcome! and thank all you guys for this discussion. In answer to @ChrisZanthens, something along those lines like the pic you uploaded would be sufficient too. I think the discussion on color coding is interesting as well.

@ms49434, Im sad to hear that hdsentinel and crystal diskmark is unavailable for freebsd. I feel like some sort of graphical benchmarking tool would be excellent to evaluate speeds in real time instead of the dd comand, Especially for people starting to use ssd's in their servers. I have once overwritten my disks badly with the wrong syntax. (Once a rank noob).
But I thank you so much as a developer for this amazing software. I will never use anything else. :.)

The smart tools in xigma that are already there work great, and I usually refer to the smart log to find what info I can use to evaluate my drives health, but i think it would make a world of difference if a user could "see" the health of the drives, and like hdsentinel, "set" a parameter for health. I usually only classify my drives as poor health when it reaches 75% health. when someone else may want to replace a drive with pending sectors at 90%. I know, it may sound weird, but perhaps some sort of extension page with graphical drive data, perhaps extending off of the venerable extended gui extension info, as this already has drive temps and status?

A great alternative for something to "display" hard drive info would be hddscan http://hddscan.com/ that would be a good option if available for freebsd. Its also a freeware.

Either way, it is a pleasure using this software for all my servers and regardless of these feature requests, I want to thank all of you for all your tireless hard work making this software better over the years.

ChrisZanthens - You the man!
JoseMR - You the da bomb! Plex is awesome!
MS49434 - You guys are, and forever will be geniuses!

Also dear friends, is no gui a hard and fast rule?

If not, smartctl does have a gui/frontend called gsmartcontrol that could be useful in an extension page or built into freebsd?

Gnome Disk Utility is another one that offers a potentially low resource gui that monitors smart values and disk activity. Its for Ubuntu/Linux so not sure if it can be made to work under freebsd.

Just wanted to offer some suggestions if this might help. Feel free to analyze or shoot them down if theyre no good :.)

Hi sleyk2010, thanks for the kind words really much appreciated.

However I'm afraid what you asking for as a request requires for a functional desktop environment under XigmaNAS, and XigmaNAS is a heavily stripped down Network Attached Storage appliance by design/purpose.

Regards

Got it. I realize that this is true. Dreams of such things are nice though. :.)
I thank you for your responses to this and wish you all a happy thanksgiving!