I’m running 2 nodes on my PC and experiencing slowdown in PC speed (Intel core i5, 8GB RAM) especially after adding second node through the toolbox. I have separate HDDs for my nodes, one 8TB which is full and added additional 10TB for the second node. in task manager I see that first HDD is often loaded almost 100% as well as System HDD. What could be eating my PCs’ power and how to improve that? maybe move storj directories from system hdd to seperate ssd?
Consider Primo Cache to offload some of the IO pressure from HDD to SSD
post picture processes of the task manager.
exeptions from antivirus for node installation folder and data folder/drive?
loglevel to error/fatal?
move dbs to ssd/usb?
automatic or manual defrag?
disabled indexing for data drives?
i don’t think it is neccecary.
how are the drives connected? what mainboard is it?
hdd are at full power for the full node its propably filewalk with defragmentation
This suggest that your system is out of memory and uses swap file on the system drive.
How much memory is used by storagenodes?
How disks are connected?
If disks are slow, storagenode could use much more RAM for buffers.
Will not help, since it doesn’t used to store customers data, unless you store customers data on the system HDD.
Ahh…
are the discs cmr or smr? Wich type/model?
Is the recommended 10% free for the full node?
IOwait can easily affect overall system performance.
i would try to move the nodes to one or two cores, so that all other cores aren’t affected by iowait.
Haven’t seen that memory would be used 100%, usually up to 60-80%.
Drives are enterprise level HGST HUS728T8TALE6L4 and Toshiba MG06ACA10TE. NO idea are they CMR or SMR?
drives connected to MB SATA slots. Disks have a reserve of 10% free space in config file.
Is it secure to add exception for data drive in Antivirus?
For storj data all the way. There are no complete files on there and they are encrypted.
Both cmr confirmed by manufacturer
Then it’s weird, why the system HDD have a usage. You may check, what processes are using the drive via Resource Monitor.
Its propably orders and logs.
Maybe orders or dbs corrupted. Check the filesizes.
I now added exceptions for Antivirus and removed indexing from second node drive. Will see how it helps.
I wonder what are db files?
The files with *.db extension. They are placed in the data location by default.
Orders.db are in the installation folder
did You dissabled filewalker?
(Windows GUI version node: in config.yaml a position to add “storage2.piece-scan-on-startup: false”)
without this, every time node restarts, example every update restarts, makes the node to scan all storj files, which is ridicules, i disabled that, and never going back.
It disables the lazy filewalker.
Sno have to choose.
Since my dbs are on usb stick or nvme, and restarts are uncommon, i switched to true.
You seems to not get it, firewalker is the root of all evil with HDD 100% consumption. a 8TB full node can scan for DAYS! Agree in one: Every SNO should figure out if it is worth wearing off his HDD like that.