I think I made a big mistake going with Unraid, am I wrong?

[MaverickMacker]

Backstory:

I currently have two NAS/Servers. Both are SuperMicro rack mounts, same processors, memory, hard drives, and running hardware raid 5. Both are running Emby, Radarr, Sonarr, SABnzbd, a database, DNS, Pihole, Syslog, S/FTP, TFTP, HyperV/Docker and DHCP. One is Ubuntu and one is Windows. In the six years they have been running, excluding long power outages, they have never been inaccessible even when a drive went bad and the new one was being rebuilt. We are moving and after thirty years as a network engineer, I no longer need a home lab (no more rack full of servers and networking equipment) work now has everything I need to work with. It’s time to downsize sort of and replace the old server.

 

My present problems:

I built a new tower server that will be on a shelf next to the media box in the av closet of the new house. After a long debate (with myself) I decided to use Unraid. Four times I had to wipe and start the configure over again (mostly my mistakes, but twice due to adding parts to the server and watching Unraid crap itself), not a good start. The issue I have right now is sort of my own making but has exposed a major issue. I moved the new server to where all the rest of the equipment I have is. During a reconfiguration of some patch cables, I tripped over the temporary power cable I had in place and the Unraid server lost power. That was three days ago. Since the server has been back on its has been doing a parity check. It’s currently at 26% saying it has 4 days and 17 hours left to go. I cannot stream anything using Emby, copy files to or from it at any speed faster than 5kbps, and none of the dockers can do anything that requires access to the array. If this was in use right now my family would be raising hell! Never mind the five kids, aunts and uncles, and parents that use the media server. My wife would be the one that would be angriest. No at me but would insist I replace the offending software/hardware with one that works. How do I move forward with Unraid if it’s going to be so unusable for days just because of a power cut? Besides going with an all-SSD setup how do speed things up? It’s not the CPU, it’s never got over 3% usage. All eight main drives are NAS 7200 rpm drives plus three cache SSDs.

[Squid]

You should post your diagnostics. 

[MaverickMacker]

mavericknas-diagnostics-20230917-1835.zip

[MaverickMacker]

1 hour ago, Squid said:

You should post your diagnostics. 

Posted the diags. Think I am going to put in a LSI 9300-8i in IT mode and replace the 8tb drives with 14s.

[trurl]

Syslog seems to indicate you have filled cache.

 

Are you trying to do the initial data load? Best if you don't try to cache anything until initial data load finished. And best to not continue to write large amounts of data to the array while parity is being checked or built.

 

2 hours ago, MaverickMacker said:

a power cut

I was going to suggest UPS, but syslog seems to indicate you have UPS though there is some problem communicating with it.

 

 

2 hours ago, MaverickMacker said:

none of the dockers can do anything

Best if you keep appdata, domains, system shares on cache or other fast pool so Docker/VM performance will not be impacted by slower parity array, and so array disks can spin down since these files are always open.

[MaverickMacker]

43 minutes ago, trurl said:

Syslog seems to indicate you have filled cache.

 

Are you trying to do the initial data load? Best if you don't try to cache anything until initial data load finished. And best to not continue to write large amounts of data to the array while parity is being checked or built.

 

I was going to suggest UPS, but syslog seems to indicate you have UPS though there is some problem communicating with it.

 

Best if you keep appdata, domains, system shares on cache or other fast pool so Docker/VM performance will not be impacted by slower parity array, and so array disks can spin down since these files are always open.

I had just moved the new server to the main UPS that is controlled by one of the old servers, but that would not have help anyway I tripped over the power cord and pulled it out of the back server. No UPS can fix that. Yes, I am trying to do the initial data load, it's only 42tb of data. The docker worked fine expect when access media files on the array. I have been meaning to move the appdata and system pack to the cache, just been concerned last time I moved them the vm and docker image files got corrupted even though both services were shut off, they seem a little fragile. At this point I am going to replace some hardware and do a factory reset, will no move those off the cache again.

Edited September 18, 2023 by MaverickMacker

[itimpi]

Your parity check speed seems much slower that I would expect - normally something between 1 and 2 GB per hour seems more normal.

 

Regardless you might want to consider installing and using the Parity Check Tuning plugin to offload parity checking to periods when the system might otherwise be idle thus trading off disruption to normal working against total duration to complete.