9yr old Rockstor Install - No UI

My NAS has been bullet proof for like 9 years, my mounts still mount in all my systems but I cant access the GUI - I get ERR_CONNECTION_REFUSED. The reason I discovered this, is I’ve gotten a few KP recently and had to manually reboot it.

Manual YUM update isn’t working - error here: gist:70b50105a05dfe5751071b688a7625d4 · GitHub

Everything looks fine on my systemctl output: gist:4bd093fb7793b342a0336fef531a7810 · GitHub

I’m kinda lost, it’s been so bullet-proof for so long I forget where to even start.

SOS - Send help lol

This might be stupid, but is the browser you are using to acces the webserver trying to force https vs http.

try manually typeing http:// before the ip or try another browser?

Yeah, wish that was it lol

Well ya gotta ask the easy stuff!

Over the years if a rare thing like that happens to me my solutions usually “screw it reinstall”

im sure there will be better advice than that though!

looks like you nginx server (which serves up the Rockstor page) is failing

Feb 15 18:50:23 rockstor.local supervisord[3510]: 2024-02-15 18:50:23,453 INFO exited: nginx (exit status 1; not expected)
Feb 15 18:50:24 rockstor.local supervisord[3510]: 2024-02-15 18:50:24,454 INFO gave up: nginx entered FATAL state, too many start retries too quickly

next step is that you can probably take a look at the nginx logs for more detail on that.

Of course, I would be remiss not to mention that maybe it’s a good time to upgrade at least to the stable 4.x version which moved off the eol’d CentOS underpinnings … but you might have other stuff on that box (considering that your yum update tries to deal with python 3.6) that this would not be quite as easy, yet.


@ScottyEdmonds Hello again.

All of what @Hooverdan said :slight_smile: .

And regarding:

keep in mind the following now long-gone issue we saw from time to time back then:

But really, there is a pressing need to get this poor machine out from the past of CentOS to our current/future in openSUSE land. We have a doc entry for that:


But often preferred to have a working system to migrate from. Let us know if it was that old empty nginx config file problem as the referenced GitHub issue has a quick confirmed work-around (in comments) to get you up-and-running again if that was the cause.

Re Kernel Panics! As you already know PSU and RAM are culprits on this front. Plus way old kernels of course :slight_smile: . And on the RAM front re have a recently updated by @Hooverdan doc section:


that has info on memory testing.

Hope that helps.


That nginx.conf being empty was the issue! Thanks @phillxnet – GUI is back up and running.

I’ll slowly work on working up the courage to migrate this to V4 lol…

I am on a 6 year old kernel lol, it’s just been so solid for so many years I’ve pretty much just went with “if it’s not broken, don’t try and fix it”.