I created this account exclusively to just ask pretty please…
IXSystems … guys please stop making operating systems. Or at the very least stop pretending to be an enterprise grade NAS OS. You guys are on par with a beta release , not a release version and most definitely not an enterprise grade release.
Please, you guys are giving IT professionals anxiety attacks all around the world, this needs to stop.
And if you insist on continuing I will pretty please ask that you remove the python powered middlewared. Then begin to do what every 1st year software engineering student learns… ERROR HANDLING!!!.
Here’s how it supposed to work: When something goes wrong … like the middlewared runs into an exception, the code is supposed to encounter a try catch block - and the catch block then tries to handle the situation.
Here’s how your system works: the python runs into a little problem and EVERYTHING crashes. The VM’s crashes, in the most annoying way possible. The filesystem remounts as RO which means that the truenas doesn’t even know to restart the VM. Jails crashes at random - some crash on one run and then others crashes at other times (randomly as far as I can tell). SMB crashes, disks detach…
You guys are gonna give me a heart attack…or an aneurysm.
My biggest regret this year: becoming as reliant as I have on your truenas OS.
Using the older FreeBSD version currently, am going to upgrade to the newer scale and if that doesn’t fix the problem I am done with TrueNAS. With all the time I have spent trying to debug your terrible OS, I could have built my own “TrueNAS” … one that can add NTFS and EXT3/4 partitions.
Oh OH and wait I forgot to mention. 2 months ago I logged onto the truenas for the first time in about 90 days. To my great surprise I found that my server was running at 100% load for over 30 days at a CPU temp of roughly 82C. Now lets ignore the power utility cost of running a server at that load level and not getting crypto currency out of it, and focus on the damage it did to my CPU because of running at 82C for over 30 days. Turns out one of my older CMR disks I had in a single disk stripe (used for small client and employees PC backups) had an I/O error and detached. The python backend gave a cryptic error of client unresponsive (or something) and after 3 x I/O errors (at different times - few minutes apart) the disk detached entirely (this was at the start of that 32 days at 82C). Then for the next 32 days it ran at 100% load (average) even with the disk detached. I don’t know if I should blame python or IXSystems developers here but it comes down to: the IXSystems devs chose to use python for a project (TrueNAS) this size…so ultimately the IXSystems devs are to blame. That is completely unacceptable for an ‘enterprise grade NAS solution’ ,
Ofc… Someone finds an issue. calls IXSystems out on it, and is attacked immediately.
It’s not as simple as migrating to another distro and you damn well know it otherwise you suck at IT.
I say IXSystems is not handling errors and you blame everyone and everything else except the poorly written python backend. Are you Donald Trump maybe? Cuz you certainly blame everyone else like him.
This is why I never ask these forums for help, I would rather dig around in logs for hours then ask a question here. It’s worse then stackoverflow
I don’t like to close off negative criticism but we do need to ensure we don’t make things personal and not to use insults.
Obviously this forum is full of people who use TrueNAS and have huge respect for the appliance and intern iX. That does not mean we agree with everything iX say and do but it does mean people will come to their defence if statements are made that they perceive to be untrue.
As always let’s keep things civil and productive. Negative feedback can be very helpful so long as it’s constructive.
Are you running your server in a professional manner?
Given what you’ve written you’re storing your client data with no redundancy and no monitoring, leaving the system for 90 days with no clue how it was running.
It’s not April 1st, so hopefully not a failed joke post, but you may be trolling. Alternatively, you appear to be skirting your own responsibilities in whatever business you’re currently employed.
Having experienced cryptic middleware issues in the past, I can sympathize with the OP. To claim that iXsystems don’t know what they’re doing seems quite hyperbole at best. To leave systems unattended for months at a time w/o alerts suggest that the OP has a few learning opportunities, regardless of whatever system he wants to switch to next.
Oh, hogwash. You posted somewhat intemperate criticisms of iX–some valid, others less so–and one user pushed back against those. And your immediate response is to play the victim, pick up your ball, and go home. Your call, I guess, but that isn’t going to result in anyone taking you at all seriously.
But he’s absolutely right about the middleware/UI’s piss-poor error handling–there’s just no excuse for a GUI to ever show a traceback to an end-user. I kind of doubt its being written in Python is particularly relevant, but this has been a problem for 15 years and hasn’t noticeably improved during that time.
middlewared. Then begin to do what every 1st year software engineering student learns… ERROR HANDLING!!!.
