noodle with crab

I recommend a Chinese movie called the wandering earth 2. Tooooo many scenes with Visual Studio LMAO

I finally understand why someone told me normally nobody knows math... OK, here's the fact: Even if you count your SSDs and mine failed 4 pieces together (I suppose you have 20 SSDs, I have 7), the failure rate is still unacceptable. You see, the AFR FIT calculated on 10k hours (mostly these drives have a power-on time around 10000 hours) = 16250, converting back to MTBF approximately 61538. However, MTBF on SSD should be 1.5M, which is approximately 1/24 of the marked MTBF. Here're some useful links to calculate the AFR FIT: https://www.ti.com/support-quality/reliability/fail-fraction-average-fail-rate-fit.html And here's the conversion equation from AFR FIT to MTBF: MTBF=(1/FIT)*10^9 Now, how likely this is "pure bad luck"? No, I don't think it's pure bad luck because it's way too high on the numbers.

I would say if you and @RONOTHAN## bought Samsung SSDs more than 2? 2.5? years ago you're probably safe, but just in case do a full disk scan with badblocks or HD Sentinel - if you have a bad drive it doesn't "show" or "behave" improperly under normal conditions, until you read that block. Then you'll notice the error count goes crazy and the disk lifespan drop down dramatically (each read-only scan takes more than 10% of the whole drive's lifespan).

970 EVOs are both M.2 NVMe. I wouldn't trust Samsung anymore, at least for consumer grade drives.

AFAIK 860 EVO doesn't fail easily. BTW these are multiple systems across different locations.

Multiple systems. I've added extra images, I have 2 in Linux with either Dell Workstation or HP Workstation, 1 in my home server (Supermicro H11SSL-i), 1 in my PC. All SSDs have firmware that are Up-to-date.

I bought a bunch of Samsung SSDs mostly this year and last year, and I've already have 4 malfunctioning units. It's not reliable. I bought approximately 7 x70 Samsung SSDs and 4 of them are having issues. Currently failed units on my hand: 2x Samsung 970 EVO plus 500G 2x Samsung 870 EVO 500G One of it failed today: And other 3 of them failed earlier this year - and I've already sent them back to Samsung for RMA. I've also heard from Reddit that 980/980 Pro are also having some victims. (See Reddit post ) Don't buy those SSDs unless you want to lose data! Here're some pictures for other 3 failed SSDs.

Is the drive reporting SMART counts(ID=0x0E, Click on Samsung Magician "S.M.A.R.T" and scroll down until you see an item with description "Media and data integrity erros")? If this number is NOT 100 (raw data) or 0x64, STOP using this drive IMMEDIATELY and contact Samsung for an RMA. You have a SSD that's malfunctioning and already have data loss. Additionally you can check for the "Available spare" , if it's not 100 then you also have a trouble. If you still have data on the drive that's important, try to copy them out without running OS on this drive. If there's nothing important on the drive, throw it away and get a fresh installed OS. There's no way to recovery those data. It's already lost.

Update: another Samsung 970 EVO Plus on my workstation in the office also showing some abnormal behavior and killing newly installed applications. It may also have similar issue. Now it's 2 out of 3 970EVO/EVO Plus dead on my hand LOL

Did a surface scan and dropped to 81% available spare.

Last year I have bought a Samsung 970 EVO 500GB (NVMe) from Amazon, and about last month I read an article talking about the evo series "media and data integrity errors" issue. So I took a look at my 970 evo on my server and I do noticed that "media and data integrity errors" is 17 since Feb. 2 this year, and since that time the "Available spare" dropped from 100% to 92% now. However, the "Data units Written" did not increase heavily and the total bytes written is only 3.8TB... For comparison, same period data units written graph The other 2 860 EVOs don't have similar issue and their "Used Reserved Block Count" is still 100% even though one of it already written 8TB data. Looks like it's better to check your Samsung SSDs and get warranty.

Looks fun on this. Hardware: Case: Supermicro SC-846 Backplane: Supermicro BPN-SAS3-846EL1 PSU: 2x Supermicro PWS-920P-SQ MB: Supermicro H11SSL-i v2.0 HS: Noctua NH-U12S CPU: EPYC 7501 RAM: Samsung 32G DDR4-2400 RECC Network 1: 2x Intel I210 LACP Network 2: Mellanox MCX354A-FCBT (40/56G VPI) RAID card: Dell PERC H730P GPU: Nvidia T1000 SSD (OS): Samsung 970 EVO 500G SSD (Cache1/2): 2x Samsung 860 EVO 500G HDD1: 7x 10T ST10000NM0086 HDD2: 1x 16T ST16000NM001G Bulk and one-time data storage HDD3: 1x 10T WD102KRYZ Backup HDD4: 1x 6T WDD6003FFBX Backup HDD5: 1x 10T HGST HDN721010ALE604 Backup HDD6: 1x 16T ST16000NM001G backup of HDD2 HDD7: 1x 4T HGST HDN726040ALE614 Backup HDD8: 1x 3T WD30EZRZ OS Backup UPS: 2x CyberPower CP1500AVRLCD Rack: Sysracks 32U 30'' network rack Power: 214W - 310W Peak 353W Software: OS: Windows Server 2019 RAID configuration: 6x10T RAID6 + 1 Hot Spare Filesystem: NTFS Capacity: 40TB/37254.0GiB (Only counting the RAID disk) Disk monitoring: HD Sentinal 5.7 Pro Administration tool: MegaRAID Storage Manager Virtualization: VMWare Workstation Pro Cache provided by PrimoCache Using Windows iSCSI target & VHDX to store actual data. Usage: Storage, some recording/transcoding of VTuber lives, internal debian-like apt source (VM), compiler, VPN server. Backup: Manual written biweekly auto backup task by copying VHDX files. Pictures/Snapshots:

This is NOT an "array". This is a snippet of JSON, so use JSON tool to read it is the best solution. Please see https://github.com/nlohmann/json and use json.hpp to parse it. This is called "deserialize" a json string.

You can get a cheap managed L2 switch and set a designated port to 10Mbps, then just connect the cable.

Fourth point makes sense. Thanks.