@scottalanmiller said in Engineering vs Administration - That's what makes Windows and FreeNAS so risky:

@stacksofplates said in Engineering vs Administration - That's what makes Windows and FreeNAS so risky:

@scottalanmiller said in Engineering vs Administration - That's what makes Windows and FreeNAS so risky:

@stacksofplates said in Engineering vs Administration - That's what makes Windows and FreeNAS so risky:

@scottalanmiller said in Engineering vs Administration - That's what makes Windows and FreeNAS so risky:

@IRJ said in Engineering vs Administration - That's what makes Windows and FreeNAS so risky:

@scottalanmiller said in Engineering vs Administration - That's what makes Windows and FreeNAS so risky:

@IRJ said in Engineering vs Administration - That's what makes Windows and FreeNAS so risky:

Admin roles are also dying with immutable infrastructure and HA. Designing a system that is immutable and highly available isn't expensive or time consuming on the cloud anymore.

But someone is still designing the initial system and someone (maybe the same person) is managing it.

Yeah so you don't have an admin here as you admit. You have an engineer designing the system and replacing the system if there is issues. It's all design and no maintenance. Maintenance is automated during build.

Not in the real world. That's a nice theory, but applies to effectively no one anywhere. In the real world, engineering almost always is a trivial effort that involves almost no time, skill or planning, and all the effort goes into years of administration that deals with that haphazard system.

That's completely false. Engineering is almost always a trivial effort......

It's completely true and I've given example after example. In the real world, engineering is generally done without planning or resources and it works enough for people to accept it. Then all the effort is hoisted onto administration. You can argue, but you can't deny that this is what 95%+ of the market does.

No you gave an example of FreeNAS and have completely ignored things like SRE where design upfront including architecture, engineering, coffee design, IaC, etc are all roles for the engineer. Immutability is vital and SREs are embedded in specific teams and only supporting that application.

Yes, but the difference is my example represents nearly the entire market. I didn't say that there weren't exceptions. But that's what they are.

Outside of F500 maybe but outside of F500 you don't normally have systems engineers and systems admins.

@matteo-nunziati said in Deciding Between Hardware and Software RAID in My FreeNAS Deployment:

@saniplastic said in Deciding Between Hardware and Software RAID in My FreeNAS Deployment:

@DustinB3403 said in Deciding Between Hardware and Software RAID in My FreeNAS Deployment:

metadata

metadata?

please explain more.

The config files where the vm is defined. Do not backup the vm disk only. I suppose he referes to this.

I backup with veeam.
whole vm files.

@uroni said in UrBackup and FreeNAS Jurassic Park Effect:

@scottalanmiller said in UrBackup and FreeNAS Jurassic Park Effect:

Does UrBackup have a process for writing to tape as well?

No, it is wholly disk based and additionally leverages advanced file system features (such as snapshots, reflinks, compression) when available.

So if we wanted to have cloud storage, we'd run a UrBackup server hosted then, rather than a local one and pointing to something like S3?

Way more discussion about this thread in this other thread: https://mangolassi.it/topic/19657/revisiting-zfs-and-freenas-in-2019

@scottalanmiller said in FreeNAS Gets Hit by the Jurassic Park Effect:

HCL

Yes, agreed!

@msff-amman-Itofficer said in FreeNAS 10 Has Been Recalled:

@scottalanmiller

True, adding the whole virtualization aspect and container support might caused this. I genuine when testing it out didnt like the new theme or steering me away from being storage server into something else.

I agree, I think that FreeNAS feels like it is drifting now. Is it storage, is it a hypervisor? Any BHyve for normal users, really? The new interface looked cool, but failed to even run when I tested it. They are a tiny company with few resource trying to do way too much. The old FreeNAS product was bad enough, I know loads of places loses data or access to data either because FreeNAS itself was glitchy or because it failed the "front loaded engineering" design principle and left customers in a tight position leading to failure.

@triple9 said in Introducing FreeNAS Corral:

@travisdh1 said in Introducing FreeNAS Corral:

@mlnews Are they going to add in a kitchen sink for the next release?

Somehow I don't see this becoming the next @scale.

not yet, for sure 😉

https://www.theregister.co.uk/2017/04/18/freenas_downgrades_latest_release_to_tech_preview/

Whoops

@scottalanmiller said in Freenas 9.2.18 server , CIFS services unable to start:

@Ghani said in Freenas 9.2.18 server , CIFS services unable to start:

Yes i agree your sounds

If the only goal is an open source file server (we call this a SAM-SD, there is a section of the forum just for that) then the most likely recommendations for an OS will be openSuse Tumbleweed or Fedora. CentOS and Ubuntu are fine choices too. FreeBSD is excellent, but less well known.

You make it sound as though the wise choice here would be to install OpenSuse etc direct to the hardware.

@Ghani said in Extend ZFS zpool Volume Size Without Any Data Loss:

@scottalanmiller

yes rsync not a backup solutions

So in a "read back" mode, this tells us that the data stored here isn't important so there isn't really anything to worry about. If you do the expansion and it causes data loss, they can't be upset as they don't see value in the data. This also tells us that the servers shouldn't be there, because if they aren't backed up and they aren't a caching system then you shouldn't have them at all.

So there is clearly something wrong.

FreeBSD does have getent.

@DustinB3403 said in FreeNAS Domain Failure on AD:

@scottalanmiller any news yet?

Nope

@tjatwood109 said in ZFS is Perfectly Safe on Hardware RAID:

@scottalanmiller

Thanks - I will proceed without using ZFS - I prefer hardware RAID.

Tim

ZFS is perfectly fine with hardware RAID, if you like ZFS' features otherwise (like zsend is nice) then there's no reason to avoid it. If you don't plan to use any unique features, then XFS is my "go to" choice by default. Very fast, very stable.

It was amazing that Scott found it so fast. I was on the Windows side of things. Inside Windows they were using the iSCSI initiator to connect to the FreeNAS. All the sudden Windows would just log a ton of iSCSI events and go down.

I looked up the events and most people resolved them by putting the iSCSI traffic on a separate NIC. This happened two days in a row at about the same time each day. I was looking at snapshot, backup, etc times when Scott found it in the FreeNAS logs.

@DustinB3403 said in Firmware Update Warning on SuperMicro from FreeNAS:

@scottalanmiller said in Firmware Update Warning on SuperMicro from FreeNAS:

@DustinB3403 said in Firmware Update Warning on SuperMicro from FreeNAS:

Why are you using FreeNAS? Every conversation ever with someone asking should I use FreeNAS or something else for my storage device, @scottalanmiller has always had the same answer. Use CentOS, or purchase a Synology.

So where did this come up that you are on a FreeNAS box?

Because we support customers. That's how customers work.

Don't get offended, I was just asking.

Don't feel bad. @scottalanmiller hasn't had his coffee yet this morning.

@dafyre said in Can You Install FreeNAS to a Single RAID 6 RAIDZ2 Array:

Who are you and what have you done with @scottalanmiller ... Actually using FreeNAS for something?

I was just about to say the same thing. Or maybe his account was hacked.

@travisdh1 said in iSCSI port w/Windows iSCSI initiator dropping:

@art_of_shred said in iSCSI port w/Windows iSCSI initiator dropping:

Is Free-NAS ever a good idea? I've only ever heard nasty things said about it.

Only if you know nothing about building and managing storage is it going to serve a purpose. BSD or Solaris without the added bits would be the preferred route to use ZFS.

And if you don't know those bits, you should not be running a SAN.

@BBigford said in Installing FreeNAS 9.3 on Scale HC3:

When you say "in a business capacity"... Do you mean any business or just certain sizes? What is the reasoning? I know you think the FreeNAS community can be very brash/vile at times based on some of your earlier posts to people asking about FN.

http://www.smbitjournal.com/2015/07/the-jurassic-park-effect/

Basically you are getting a highly stateful system where support is critical but in a crippled manner compared to just using FreeBSD. You are getting something easy to set up but difficult to support. If anything goes wrong you are in very tough shape. And updates come a bit behind. So you have a number of small issues that all add up to a not very business friendly product.

Looping back to this, in the past month I've worked with three different companies that all experienced significant data loss or downtime because of their choice of FreeNAS. Two suffered from not having front loaded their engineering and had an inability to support their servers during routine operations and caused major outages because of it along with significant cost for repairs, and one company that lost its data because of unnecessary bugs in the FreeNAS GUI code that would have been avoided has they been simply on FreeBSD.

Additionally this past week FreeNAS 10 "Coral" was demonstrated to be so incredibly unstable a month after being released that they had to recall the release and revert to a "beta" status indefinitely. For a trivial end user application this would be bad, for a critical storage infrastructure component on which companies need to have rock solid faith, it's unthinkable.

Literally just got off of the phone with someone who had a FreeNAS bug cause a system to become useless, just like the one case that happened today. But instead of it being the JPE encouraging a mistake, the same problem happened through a GUI bug. One of the big risks that the JPE introduces is that the GUI is all "extra" points of failure and has nowhere near the testing of the standard OS tools. So a little big can cause a lot of damage, as it did. The entire SAN had to be replaced due to a small bug in the FreeNAS interface.