GRRRRRRR

The H240 has (8 physical links x 12Gb/s) also from the PDF Mix-and-match SAS and SATA hard drives; deploy drive technology as needed Looks like a firmware update is needed to support the SSDs better. What are these SSDs that this HBA can not saturate, but the Asmedia ASM1166 can ?

Recently I like WD dual actuators and optinand and the metadata on the logic board. I would also like to see the metadata backed up on the HBA and on the host as well. I am also perfectly fine with fully saturating the available rack space with different disks, configured in a way that reduces stress, vibrations, and so on creates a good environment where the cheapest drives are not really expected to fail or cause problems during replacing. I can also create partitions that skip the beginning of a drive if I am not sure how it was used previously so there is a less chance of defects near metadata area or overused area. Sometimes I can repair PCB I don't really count that as drive failure if I can repair it cheaply by swapping burned resistor capacitor or diode.

In this case what you can do is to passthrough the whole PCI-e device including all it's subcomponents - all four ports. So basically all four ethernets that are on the same PCI-e PCB. ninguno disponible - none available Boot ESXi on the machine and check if the passthrough works on ESXI if it works on ESXi and not on unRAID then ask the unRAID dudes in General support however this time use translators on your end to speed up the process.

PCIe passthrough may not be available, due to some reasons something having to do with sharing memory or something. Is that one PCI-e device with 4 ethernet ports ? In this case you can try adding another PCI-e device for that passthrough, that you want to do. Personally I like minimal processing. If passthrough will save CPU cycles, I want to do it. Sometimes I do testing, and in the end use separate device in production.

Investing in IT since I can remember myself, can't get any profit whatsoever. Salary is high AF still not worth it at all and in the negative. No solution in sight other than building everything by my own two hands and defending everything with my own fists. Living a prison life to the problems in IT. Corporations produce crap products that do not work. Problems are multiplying exponentially. Everything is breaking all the time software hardware. Hardware is not performing. Irresponsible liars everywhere. Stupidity and chaos everywhere. Everything is being hacked. This is the reason and when I see someone lying, you know what is coming their way. Beyond Punishment. ☠️

I noticed the weird 260+ character directory when I was doing data migrate with MultiCommander. The App asked whether to keep the 260+ long character name.. I was annoyed AF and didnt screeshot that MF. I tried a bunch of tools to look at the filesystem and even r-studio, didn't find anything on the filesystem. It could have been a problem with the app or a problem with a DLL module in the system. Since them I killed 5-6 unused low-trust modules and drivers. Since this is Win you can't scratch your balls without running some low trust app that leaves behind some crap. The other problem with the system is that when I switch to an App such as MultiCommander, explorer.exe restarts with mui xml error in the eventlog. The 4x security systems detect nothing suspicious. No data loss or corruption thankfully. No client or company or sensitive data leak either, I think. that is protected AF in some weird ways.

Hi, I saw on one of my NTFS a folder F:\_Company\Windows\Windows\Windows\Windows\Windows\Windows\Windows\Windows\Windows\Windows\Windows\Windows\Windows\\< some Chinese hieroglyphs> But then it disappeared without a trace. Maybe because I turned On a bunch security systems. How do I Forensics this thing ? Thank you.

What? Why? I don't see why not to put the correct quality towel, infront of the dust intakes ? Everything said and done, looks like the best dust filter for the task. Can even put a cool t-shirt with some picture infront of that. I didn't understand your 'metaphor' ? Or am too lazy to research that part of the movie. Is it a sex joke? There are also airfilter materials for different purposes such as for vacuum cleaners however those are not the right size usually and one time use as well as will look worse. So to get a cool looking machine you are limited to specific things to put infront of it.

Eliminate writes to the USB limit that to mostly reads. Not 100% sure how exactly with unRAID. Remove the plastic around the USB stick and mount it inside the machine, in a way that facilitates airflow around the USB Stick and it's components. Put some hot glue or capton tape or high quality industrial adhesive tape, to avoid something getting loose and loose electrical connections. If it fails again, obviously they were made to fail. Hit them hard with right to repair action the USB manufacturers/brand. Meanwhile the possibility is to update the USB flash controller firmware. It happens when a stick is powred On for long periods, it drops out of the system. I believe the correct fix for this is to get the right MPTool and firmware and update and "remanufacture" the USB flash drive (there is format process involved with badsector avoidance logic).

Actually, it should also work to put cotton cloth / towel Infront of the machine this way you can have choice of color etc. and wash it with the laundry. The towel will also absorb some of the noise. The trick is to have a big intake area.

There are also 'disposable kitchen towels' aka 'absorbent rolls' for home and industry. They are pass air you can use for cheap dust filters. Those would be the products with more cotton than paper.

I think it's a good system to run the dockers and the VMs on unRAID and that it should be the right environment to become more stable and better. Better than Proxmox. If Lime Tech expand a little.

Well if it works sufficiently stable? It can always be optimized later. Docker is not optimal I don't know why it became so popular. Many want to replace it but I don't remember wiith what exactly LXC ... ? Need to doublecheck that. Kubernetes will be moving away from docker containers is what I know however each system can prolong that 20 years if they waynt unless we instigate some fights or something.

Don't switch stuff the outputs have a specific order. You can also try to check cooling on the LSI and update the firmware on it. During the firmware update you can reset to defaults, this can often help.

The LSI cards can connect disks in JBOD mode that's the same like IT mote maybe with different "queue depth" Also if you want to use the cache on the LSI you will need to at least RAID0 a disk and maybe a battery/supercap. What do you want to do with it?

Test with 1 HDD in that backplane. The backplane can have firmware updates. Supermicro are good at firmware updates, you can also ask for those over e-mail and hunt them on their websies. You can also try simple checks like voltages dust, metal dust particles, loose cables connectors. Wrong ports and unplugged stuff - this can happen. Keep doublechecking things.

I didn't read the post but I fully agree. As a matter of fact a personal data storage system with VM container and media capabilities, must anticipate the end to end life of data and workloads, and Storage QoS and power management dimensions. How is the unRAID or Linux storage systems documented in terms of high level design logic ? Where are the diagrams?

Yes I wanted to make the same post. I just want to use junk storage in high iops Storage Spaces without protection, and protect them with raid1 btrfs raidz1 or similar. And I want to run the containers and the VMs on unRAID. I want to go fast cheap. I also want to turn off the HDDs so yes READ CACHE everything and spin up HDDs rarely. Are you aware of a Linux filesystem that can spread the IOPS across different media, like how Storage Spaces does?

So you think your storage is the shizzle. But have you considered what Storage Spaces with ReFS is capable of doing 😈 Here is the virtualization layering: 1. unRAUD host passthrough virtual SCSI to 1x or 2x Win Dev build VMs or tiny11s 1.1. Per VM passthrough a fast nvme - same ot each VM or split it by passing virtual scsi or whatever 1.2. Per VM plus any other whatever disks passed through same to each VM 2. On The VMs do a Storage Spaces without protection on the available disks and ReFS 😈 ----> What does this do? This will create auto-tiered NVME cache which' capacity is part of the Storage Space. Mind blown, pants pooped. Writes go fast then ReFS rebalances them later. The MS code writes parallel with high iops so combining may different storage can work. The unRAID Array can not do this with it's simple XOR-ness. 3. Add a RAID1 layer somewhere either on the VM or on the host, depends on which one will be with the lesser overhead 4. Enjoy the best balance between performance and protection in the known universe. Congrats. You have broken the matrix. For sure this can be optimized somehow.

Good job. I also use wet wipes, when they are dry, to catch dust infront of air filters and fans.

How does unRAID handle Storage Tiering and Storage QoS ? Now and in the Future ? Load balancing across mixed media ? I would like to build Universal Storage servers this makes the most sense all things said and done. I would like to do Storage Tiering with all types of Storage currently available to me wherever I can source them. With and without ECC, power loss protection, enterprise or 'noname' - I put in the whole thing. I think the best unRAID endgame is to have auto-tier-discovery in Software and some auto-config. 1. RAM which is or is not backed by UPS-ing the storage server 2. nvdimm or battery backed RAM 3. low latency high iops pci-e / nvme such optane or some nvdimms 4. pci-e or datacenter fabric NVMeOF and the big HPC and supercomputing stuff - this is realistic to support because this is on ebay all the time 3 years within release or even if its 13 years after release I still put it in the storage / VM machines I build 5. low latency pci-e nvme such as some newer models and some enterprise models, with or without power loss protection .. .. .. N. New HDDs N. Old HDDs N. SMR HDDs N. HDDs beyond their service life N. Some cheap flash N. Perfocard N. carving on cave wall What do you think and what can we expect and what can we do when we mix all these?

Nothing is cheap end-to-end. What is this flea-net that stops all the dust entering? Is it a finer net ? I disagree with the flea-net. I can agree with hepa however even these are up to 0.3 uM

That's pretty cool. How did you do that? How do you use it? With what containers, apps and stuff?

I wouldn't do it. Use a system that treats the hardware in a more as designed matter, with the trim working. Unless the drive manufacturer specifically officially in writing supports operation without trim, and with built-in trim, then run the drive with the bult in trim. Otherwise simply say No. Why unRAID does not have trim in the Array ? If this is the case just don't use Array. Use them with ZFS or directly. The consumer crap can also do the writes you just need to use the biggest models and write little to them. As well as write to RAM. Just use UPS and write to RAM this is more cost efficient.

So we should hit them all with Right to Repair action? We must fight for the right to run Linux on everything !!!!