“Rapid development” strikes again. FreeNAS has gone down that route and the results are quite predictable. The last 2 releases since their announcement of their new dev model have had at least one regression each..u3 had two of them..including an error they introduced into ZFS itself!. If anything at all..DO NOT MESS WITH ZFS!. It turns out that the current release of FreeNAS has a couple of additional errors in their middleware that are preventing some basic things from working..things that have either worked for years or from the very beginning:

replication is broken in 11.3u4.1(the latest version) https://jira.ixsystems.com/browse/NAS-106682?filter=-2

This bug has completely broken ZFS replication in the FreeNAS interface.

seriously? This is built into ZFS itself…and you borked your own interface where you cannot activate this any longer? Something that’s been possible from the beginning?

seriously? This is built into ZFS itself…and you borked your own interface where you cannot activate this any longer? Something that’s been possible from the beginning?

Second:

The ability to use rclone to upload to Backblaze B2. This used to work..but in the 11.4 version..again they borked their middleware so that a file over 2TB is not able to be uploaded to B2:

https://jira.ixsystems.com/browse/NAS-107032

So this bug is because when the middleware passes the command to rclone they forgot to increase the upload cutoff command to match the chunk size. Since B2 has a 10000 chunk limit..200 megs at 10k chunks in 2TB. That’s why I’ve been having multiple failures with B2 now on multiple machines..including an 8TB upload from a remote network in California.

The final one..unable to delete file based ISCSI extents:

https://jira.ixsystems.com/browse/NAS-107133

It’s one thing if the web interface fails…but now even at the command line…the file’s cannot be deleted because supposedly they do not exist even though the web interface created them…and for the most part the ISCSI connections..except one..work. I had to delete one of the ISCSI connections because ISCSI for some reason refused to run..when i delete that connection from FreeNAS iscsi works..but now the file is…stuck.

Why am I bringing this up? right now the network in Cali is being held up and backups are not working correctly because of the buggy mess that FreeNAS has become. It’s going to mean either we have to put out more $$$ to replace the official FreeNAS mini with a QNAP OR we are going to loose the job after a month. We have a TON of hours that we will not get paid for because of two of the above bugs that are causing everything we are trying to do to fail.

The replication bug also caused me to have issues at my radio station client(along with a Unifi DHCP/DNS bug that took down the network briefly..but that’s another post) so that i am unable to install the r310 into that building. I also would like to be able to do a replication from my 520 to the 420..but right now that’s not doable via the interface(yes i know the cli is still there..but if I wanted to do that all in the cli i would just run straight BSD instead of BSD with an interface on top).

Right now I have three projects active that are all halted by middlware bugs on FreeNAS directly attributive to the latest switch to the raging quick dev model. I have also informed IXsystems(via my rep) that as of right now the 40K project for the California school system is indefinitely suspended. That got their attention and they are going to setup a meeting next week with us…and I am going to be very blunt:

I do NOT care for Linux…at all…(Linux svchost..aka SystemD)..has taken over..and it’s garbage. Right now I am also suspending ALL FreeNAS projects until further notice because if this is the quality of software we can expect when 12.0 comes along(as they are going to have Free/TrueNAS on the same codebase then I want no part of it. If a simple 4k project is being torpedoed because of defects in your free product…how can we trust your products to be not full of the same defects when the code bases are merged? If a simple 4k FreeNAS backup target project cannot be successfully accomplished…how can we trust a much more critical 40k project to succeed?