Hello,
I recently upgraded to Scale.
I’ve noticed that my Boot drive gets a lot more activity than I expect, especially compared to what I recall it getting when on CORE. Is this normal in Scale? Should I be concerned or just let it do its thing? My boot pool isn’t mirrored. The system dataset is on a separate mirrored pool of NVMEs. The boot pool is a consumer grade 500 GB Samsung SATA SSD.
Dragonfish-24.04.0
Thanks!
Is you log being spammed with something aggressively?
Hi @winnielinnie,
Any suggestions on which log I should check? /var/log/syslog is pretty quiet. Noticed that /var/log/netdata/error.log is getting hit with this every second:
However, that doesn’t look like enough to generate the writes I’m seeing in the graph.
Here’s log files with writes in the last minute:
Per htop I/O screen, middlewared is going nuts with reads, not so much writes.
However, I’m not sure which drive it is writing to.
Middlewared has calmed down. Not sure what it was doing. Boot pool is still getting hit pretty hard.
Probably why I had to add a heatsink to my boot m.2 after upgrading to scale…
Just by gauging the users from the old forum (and continuing in here), SCALE seems to be more “chatty” than Core.
I love Core. 
I guess it just must be normal! I was used to Core barely writing at all… so it just alarmed me at first. I guess I’ll just let it do its thing. 
Looks like the 870 EVO has a lifespan of about 300 TBW… so I guess it should be good for a while still. at least it’s not a USB thumb drive! 
If it helps, here is my Scale boot drive reporting data:
I am on Cobia though, too early to migrate to Dragonfish for my tastes. All the logs go to the boot drive since Cobia. didn’t use to.
I figured it out. For some reason, my system was swapping, even though I have ~240gb of ram. I ran swapoff -a and then disk usage slowed way down. Bonus- it seems faster too. Not sure why it was using swap. You can tell in the graph below exactly when I turned off swap.
edit:
adding another graph to show when swap turned off:
I’m feeling quite confident that swap was hitting the boot drive for some reason. I dunno why, but turning it off made my boot drive shut up.
I added this as an init script in the UI so it applies next boot. Good idea?
Dragonfish writes logs and audit entries to the boot device.
Your chart with it off looks basically like my Cobia one, which I consider “normal”.
I wish you had posted the memory reporting.
So, it appears you got down to (at least) 12GB ram, could be lower given the interval it collects. Would have been interesting to see swap reports at the time. I’m not surprised there was some swap used given how memory management works in Linux. I AM surprised you would be thrashing in and out. What kind of drives are the swap on?
cat /proc/mdstat
SATA connected SSD, 500 GB Samsung 870 EVO.
root@storage01[~]# cat /proc/mdstat
Personalities : [linear] [multipath] [raid0] [raid1] [raid6] [raid5] [raid4] [raid10]
unused devices: <none>
Also see this thread. I’ve posted a fair amount here after discovering that swap was being “chatty” on my boot disk.
Very slow WebUI - Login, Apps, etc. SCALE Dragonfish RC - TrueNAS General - TrueNAS Community Forums
Surprised then that your swap was so utilized. That should not be the case but stopping clearly beneficial here. Just surprised your UI was slow given SSD swap, must have really been thrashing!