MurtaghsNAS
Dabbler
- Joined
- Jul 21, 2021
- Messages
- 17
I am in the process of setting up my first TrueNAS Scale server, and thought I would record some of thoughts and experiences from doing this for the first time. I doubt it will have much value beyond UX (user experience) designers and other newbies previewing the process, but recording it is important because soon I will be experienced, and my perspective will be different.
Who Am I? I am a fairly experienced IT pro. I have filled a lot of roles professionally, from validation to support to systems administration, in a variety of organizations. I have never been a hardcore storage administrator, but you have to deal with storage at nearly every level of IT so you pick stuff up.
Why am I using TrueNAS Scale? I currently have a 4-bay QNAP system that I have been using for 5+years. It is a nice system, but between it approaching EOL and a need to massively upgrade my storage requirements to support digitizing my movie collection, it was time to consider upgrades. I had heard of FreeNAS/TrueNAS, and originally investigated them to be run as a VM. My timing however was excellent, and I discovered Scale. For my needs it offered a cleaner, better architected solution than my original plan of a VM server paired with a storage server.
First Install: After an overly dramatic hardware procurement process (let's just say a mere part of the story is FedEx taking 5 days to move a package 20 miles), I was finally to install Scale. I made my install disk, ran it, and the install just worked. In 10 minutes, the system was up and running happily. I had planned on at least three install attempts: one a disposable exploration, one a "oops, wrong way", and a final. But it all worked well the first try. This is not the typical for a complex piece of software. I mean, I have installed MythTV a dozen different times and I still don't feel confident I can do it right the first time. So Kudos to the TrueNAS team!
Early thoughts:
- The one thing I find missing compared to my QNAP is their file management utility, File Station. I really like having a one stop location to create directory structures, move files from one dataset to another, etc. It seems like something that could be filled with a custom app (or even a curated stock choice) added to the official apps.
- Now that Scale has a Linux base, supporting a couple of the more common file formats other than zfs could be advantageous. My case study is this: I have a m2 disk to house my VM operations on. I know it is a high-risk-of-loss configuration, and accept that risk. When I went to set up that pool, my only option was a striped configuration. I rather suspect I have hit upon a bug with that option (I was expecting zraid0) but that is beside the point. ZFS is not an ideal choice for any single drive configuration. An ext4 formatting would have been a better choice.
- I think going rootless might be a security improvement. Having the knowledge of the name of an all-powerful account on every Scale server is an advantage for someone trying to hack the servers. A comparison is an Ubuntu installation. With Ubuntu's rootless design, before a bad actor can compromise an administrator account, they have to identify one. It is a step that can be over come for sure, but every step you can slow the bad actors down is a win.
- One thing that I am struggling with currently is catching up with the tribal knowledge. This is true of most communities of specialists, but the exacting way language is used in TrueNAS and ZFS is more complex than normal. I'm sure once I have marinated in it for a bit things like ZVOLS and L2ARC will become clearer, but it takes time. Please when you are dealing with the newbies that Scale is going to attract, remember we barely know the lingo, much less the arcana.
- An addition to the forums here that would be nice to have is to pin a thread that details how to log a bug. Where do you go? What logs are expected and how do you get them. What is a best practices form explaining the bug that the developers can quickly triage. I think a good statement of expectations from leadership would generate more, higher quality bug reports.
As I populate my server, and proceed onward towards challenges such as transferring my VMs, I just wanted to log these first impressions. I am impressed so far. For a complex project still in beta, Scale is is really good shape. There are things that can and will be improved, but the team is doing a very good job and should feel proud of what they have done so far.
Who Am I? I am a fairly experienced IT pro. I have filled a lot of roles professionally, from validation to support to systems administration, in a variety of organizations. I have never been a hardcore storage administrator, but you have to deal with storage at nearly every level of IT so you pick stuff up.
Why am I using TrueNAS Scale? I currently have a 4-bay QNAP system that I have been using for 5+years. It is a nice system, but between it approaching EOL and a need to massively upgrade my storage requirements to support digitizing my movie collection, it was time to consider upgrades. I had heard of FreeNAS/TrueNAS, and originally investigated them to be run as a VM. My timing however was excellent, and I discovered Scale. For my needs it offered a cleaner, better architected solution than my original plan of a VM server paired with a storage server.
First Install: After an overly dramatic hardware procurement process (let's just say a mere part of the story is FedEx taking 5 days to move a package 20 miles), I was finally to install Scale. I made my install disk, ran it, and the install just worked. In 10 minutes, the system was up and running happily. I had planned on at least three install attempts: one a disposable exploration, one a "oops, wrong way", and a final. But it all worked well the first try. This is not the typical for a complex piece of software. I mean, I have installed MythTV a dozen different times and I still don't feel confident I can do it right the first time. So Kudos to the TrueNAS team!
Early thoughts:
- The one thing I find missing compared to my QNAP is their file management utility, File Station. I really like having a one stop location to create directory structures, move files from one dataset to another, etc. It seems like something that could be filled with a custom app (or even a curated stock choice) added to the official apps.
- Now that Scale has a Linux base, supporting a couple of the more common file formats other than zfs could be advantageous. My case study is this: I have a m2 disk to house my VM operations on. I know it is a high-risk-of-loss configuration, and accept that risk. When I went to set up that pool, my only option was a striped configuration. I rather suspect I have hit upon a bug with that option (I was expecting zraid0) but that is beside the point. ZFS is not an ideal choice for any single drive configuration. An ext4 formatting would have been a better choice.
- I think going rootless might be a security improvement. Having the knowledge of the name of an all-powerful account on every Scale server is an advantage for someone trying to hack the servers. A comparison is an Ubuntu installation. With Ubuntu's rootless design, before a bad actor can compromise an administrator account, they have to identify one. It is a step that can be over come for sure, but every step you can slow the bad actors down is a win.
- One thing that I am struggling with currently is catching up with the tribal knowledge. This is true of most communities of specialists, but the exacting way language is used in TrueNAS and ZFS is more complex than normal. I'm sure once I have marinated in it for a bit things like ZVOLS and L2ARC will become clearer, but it takes time. Please when you are dealing with the newbies that Scale is going to attract, remember we barely know the lingo, much less the arcana.
- An addition to the forums here that would be nice to have is to pin a thread that details how to log a bug. Where do you go? What logs are expected and how do you get them. What is a best practices form explaining the bug that the developers can quickly triage. I think a good statement of expectations from leadership would generate more, higher quality bug reports.
As I populate my server, and proceed onward towards challenges such as transferring my VMs, I just wanted to log these first impressions. I am impressed so far. For a complex project still in beta, Scale is is really good shape. There are things that can and will be improved, but the team is doing a very good job and should feel proud of what they have done so far.
