Should I upgrade to 5.0.6?

[Raident]

First of all, I'm very happy with 5.0.5 and have been for a long time now.

 

With 6.x, however, LimeTech seems to have shifted focus from providing a NAS to providing a media server. Given that my unRAID installation runs on top of ESXi and I already have dedicated VMs covering most of the new features advertised in 6.x, this is fundamentally at odds with my use case. And even if I ignore all that stuff, there's still the need to go through validation all over again, so either way it would be in my best interests not to upgrade to 6.x.

 

That leaves 5.0.6 - I'm torn between upgrading to what is essentially the final 5.0.x service release and being on the same version as everyone else for support reasons as well as having the opportunity to do a clean install and wipe out all the plugins that I tried but ended up not using on the one hand, but on the other hand, there's the old adage "don't fix what ain't broke"...

[abs0lut.zer0]

I'm sure i am definitely the LAST person to try and give advice but:

 

Changes from 5.0.5 to 5.0.6

---------------------------

- emhttp: explicitly set "sec=sys" in exports file for NFS public and secure mode share export

- emhttp: fixed bug where blank value for cache 'Min free space' would cause shfs terminate

- emhttp: fixed bug where changing SMB export setting in Active Directory mode causes crash

- slack: update bash-4.1.014-i486-1_slack13.1

 

Known issues in this release

----------------------------

- slack: the time setting "(UTC+10:00) Brisbane" is wrong: Brisbane does not use daylight savings time like (UTC+10:00) Camberra, Melbourne, Sydney.

 

if nothing here is for you or affecting your install

 

then as you said "don't fix what ain't broke"

 

but i don't think going to 5.0.6 should break anything on your install

going to 6 a very different story

[CHBMB]

Personally if the only reason you're upgrading is for support reasons then I would either leave yours alone or bite the bullet and upgrade to V6.

 

V5 support has become increasingly difficult as very few if any people who provide the support actually run V5 any more.  The general trend is that the person is advised to upgrade to V6.

 

 

[landS]

The exception to moving from v5 to v6 is if you have slower hardware.  My D525 atom server has severely limited ssd cache write speed ... And has lost the ability to function smoothly under load from multiple users that it didn't break a sweat with in v5.

...

If your hardware is up to snuff.... I think garycase recommends a min passmark CPU of 5,000 for nas use..... Then v6 is phenomenal!

[trurl]

And there have been significant improvements and additions that make V6 NAS functionality better. SMART and other notifications for example. I have seen a number of examples where a user literally "set it and forget it" on V5 or older and when they finally discover it is broken, they have multiple problems that make it difficult to recover from. They may have been running for some time with one of the disks emulated for example and didn't even know it. Notifications are critical to NAS functionality in my opinion.

[Raident]

I'm sure i am definitely the LAST person to try and give advice but:

 

Changes from 5.0.5 to 5.0.6

---------------------------

- emhttp: explicitly set "sec=sys" in exports file for NFS public and secure mode share export

- emhttp: fixed bug where blank value for cache 'Min free space' would cause shfs terminate

- emhttp: fixed bug where changing SMB export setting in Active Directory mode causes crash

- slack: update bash-4.1.014-i486-1_slack13.1

 

Known issues in this release

----------------------------

- slack: the time setting "(UTC+10:00) Brisbane" is wrong: Brisbane does not use daylight savings time like (UTC+10:00) Camberra, Melbourne, Sydney.

 

if nothing here is for you or affecting your install

 

then as you said "don't fix what ain't broke"

 

but i don't think going to 5.0.6 should break anything on your install

going to 6 a very different story

 

Pretty much - it's the kind of situation where my better judging is trying to hold back my urge to tinker

 

Personally if the only reason you're upgrading is for support reasons then I would either leave yours alone or bite the bullet and upgrade to V6.

 

V5 support has become increasingly difficult as very few if any people who provide the support actually run V5 any more.  The general trend is that the person is advised to upgrade to V6.

 

Thanks, this is good to know.

 

The exception to moving from v5 to v6 is if you have slower hardware.  My D525 atom server has severely limited ssd cache write speed ... And has lost the ability to function smoothly under load from multiple users that it didn't break a sweat with in v5.

...

If your hardware is up to snuff.... I think garycase recommends a min passmark CPU of 5,000 for nas use..... Then v6 is phenomenal!

 

And there have been significant improvements and additions that make V6 NAS functionality better. SMART and other notifications for example. I have seen a number of examples where a user literally "set it and forget it" on V5 or older and when they finally discover it is broken, they have multiple problems that make it difficult to recover from. They may have been running for some time with one of the disks emulated for example and didn't even know it. Notifications are critical to NAS functionality in my opinion.

 

Unless if

 

a) LimeTech officially supports my upgrade scenario (they don't as far as I can tell - nothing is mentioned in the upgrade guide and it seems unRAID has moved from happily running on top of a hypervisor in V5 to providing a KVM hypervisor in V6 and expecting everyone to rebuild their entire infrastructure on top their new architecture), or

 

b) Someone with a very similar configuration to mine has already gone through the whole upgrade process already and is willing to share their experiences and answer a bunch of questions (a few questions to start off: Is there a VMware Tools plugin for V6? Is VT-d passthrough of the SATA controller confirmed to work with 64-bit guests in ESXi 6.0? What is the process to install V6 to a VMDK file as opposed to a USB stick? Does V6 have ESXi drivers?) thus greatly simplifying my own validation process,

 

It's simply not going to happen regardless of the benefits V6 brings. I don't have the time to go through the whole trial-and-error route of creating a custom installation/upgrade process.

[CHBMB]

i don't run Unraid as a guest on ESXI, but it's definitely do-able...

 

I do run Unraid as guest on an Unraid host though and I believe the process is similar.

 

There's a whole subforum dedicated to it here.

 

Whether you'll be able to answer those queries you have I don't know as I've never used ESXI...

[Raident]

i don't run Unraid as a guest on ESXI, but it's definitely do-able...

 

I do run Unraid as guest on an Unraid host though and I believe the process is similar.

 

There's a whole subforum dedicated to it here.

 

Whether you'll be able to answer those queries you have I don't know as I've never used ESXI...

 

I'm sure it's do-able It's just that there's a zillion compatibility-related factors that need to be re-checked (if someone else hasn't already done it then that means I need to spend the time re-running all the tests I initially ran before picking unRAID V5 as my NAS solution in the first place - and that took about 2-3 months worth of weekends), and on top of that, now that I'm actually using the system there's the added complication of "will running this test destroy my data?"

 

Given the projected headaches and the huge risk, I can't really justify it for the rather meagre benefits. I'd rather be spending my weekends going on ski trips or watching a movie or something instead. I suppose it going over and asking never hurts, though...

[garycase]

FWIW, I didn't bother upgrading my v5.0.5 server to 5.0.6 until there was a power outage that forced the server to shut down (after about a year or uptime).    Since I had to reboot it anyway, I went ahead and did the v5.0.6 upgrade, which was totally painless (just copying the two key files).

 

But as I understand it, doing it under ESXi is a bit more involved ... so I'd be inclined to simply leave well enough alone unless there's some issue you're hoping to resolve => and I gather from your initial comment that this is not the case:

 

... I'm very happy with 5.0.5 and have been for a long time now.

 

 

Given that, I'd be inclined to just leave everything alone, for the same reason you noted at the end:

...  the old adage "don't fix what ain't broke"...

[RobJ]

For completeness, I'll note that there have been a number of security vulnerabilities fixed, only available in the latest v6 releases.  If you are sufficiently technical to maintain and secure your own network, then that's still not a reason to upgrade, but you should probably consider your system like an XP or Win98 system.  It will require a more technical person to maintain, and all networked devices will have to be more closely monitored.  An infected machine on your network can attack your unRAID server.

[Raident]

For completeness, I'll note that there have been a number of security vulnerabilities fixed, only available in the latest v6 releases.  If you are sufficiently technical to maintain and secure your own network, then that's still not a reason to upgrade, but you should probably consider your system like an XP or Win98 system.  It will require a more technical person to maintain, and all networked devices will have to be more closely monitored.  An infected machine on your network can attack your unRAID server.

 

Well, LimeTech was very upfront about V5 being inherently insecure - I believe the official documentation mentioned that it should only be run behind a firewall, never to be exposed to the public internet. I still remember their response to Heartbleed was something along the lines of "V5 uses insecure Telnet, so there's no impact".