The IO operation at logical block address "x" was retried. Cheap HP SSD errors in Event Viewer and poor performance.

[criccio]

Getting this out of the way:
Windows 10 Home 1909
2700x @ Stock
Asrock AB350 iTX
16GB DDR4
2080 Super

My storage situation is a bit of a mess and I really, really need to consolidate.
OS/C: Drive: 500GB Samsung 970 Evo NVMe
D: Drive: Very cheap HP S700 Pro 256GB SATA <----- Culprit!
E: Drive: Old Intel 520 Series 120GB
F: Drive: Ancient 7200RPM 2TB Seagate HDD
G: Drive: Samsung 840 Evo 120GB

Been using this HP drive for a couple games for a few months just fine. However in the last few days I installed Minecraft Dungeons and Warzone on and began to notice some serious issues where, what I now know is the drive just not responding for large amounts of time. This results in Warzone essentially hard locking and some hilarious texture missing in MCD.

Thinking there was an actual issue with the drive, I promptly did a full wipe and format with diskpart which.. took longer than it should have. I then tried just copying a large file to it and it would stall at time then eventually resume. Sadly there is no dedicated software from HP, at least that i'm aware for doing any diagnostic's on this drive, however CrystalDiskInfo doesn't seem to show anything specifically wrong:



Either way, whenever this happens, Event Viewer is absolutely FLOODED with the following errors:

The IO Operation at logical block address 0x8018 for Disk 0 (PDO name: \Device\00000037) was retried.

Disk 0 according to diskpart is this HP SSD. For each error, the "0x8018" is something different, which makes sense. Check Disk on the drive comes back with no issues as well when running it manually on the D: drive from the command prompt.

Googling isn't getting me anywhere, really. The best i'm coming up with is it isn't necessarily a hardware fault, however i'm finding that hard to believe. Before I went and started trying some of the things listed here, I wanted to ask around here first. Could this just be a cable? As it turns out, i'm swapping cases later tonight or tomorrow so I'll probably do that.

http://www.pwrusr.com/system-admini...at-logical-block-address-for-disk-was-retried

 

[sinisterDei]

It could be the cable, or the port on the mobo. Might as well try both before giving up on the SSD.

Upside is that it's a super cheap SSD and it sounds like it's still working technically, just having to retry operations to get them to work. If you have data on there you care about.... ABANDON SHIP!

 

[criccio]

It could be the cable, or the port on the mobo. Might as well try both before giving up on the SSD.

Upside is that it's a super cheap SSD and it sounds like it's still working technically, just having to retry operations to get them to work. If you have data on there you care about.... ABANDON SHIP!

Thanks, that's the plan. Luckily it was just a few games so I wiped the drive as a first step so no issues with data loss, thankfully!

 

[likeman]

8.5.2 is current version

can you post it in decimal (Function > adv > Raw values > pick Dec) it remembers once you picked it (unless you update diskinfo then it seems to res it back to Hex)

the drive has had a lot of writes on it (316TB Written) witch is likely why your having problems and the SSd is not handling it very well, most SSDs at that point would of been at 0% expected life left at that point, but your SATA CRC errors might be high unsure due to it been in Hex (sata cable is likely the problem)

 

[criccio]

8.5.2 is current version

can you post it in decimal (Function > adv > Raw values > pick Dec) it remembers once you picked it (unless you update diskinfo then it seems to res it back to Hex)

the drive has had a lot of writes on it (316TB Written) witch is likely why your having problems and the SSd is not handling it very well, most SSDs at that point would of been at 0% expected life left at that point, but your SATA CRC errors might be high unsure due to it been in Hex (sata cable is likely the problem)

No problem. That's a really good catch, i'm blown away by those host writes. I'm actually completely unsure what was even writing to it for that long seeing as I have a much older Intel SSD in the same system with only 30TB written. You're really probably onto something there.

 

[likeman]

No problem. That's a really good catch, i'm blown away by those host writes. I'm actually completely unsure what was even writing to it for that long seeing as I have a much older Intel SSD in the same system with only 30TB written. You're really probably onto something there.

the retired block count is extremely high (x10000) i would forget about using that drive any more (i would imagine it's already burned out the spare area)

the ID >AD< is likely Pending or total sector problems (as it's about 41 more then the relocated one)

you have close to 2x past the expected Write Endurance
https://www.anandtech.com/show/11790/the-hp-s700-and-s700-pro-ssd-review

 

[Maxx]

I wish I could tell what values AD and B0 actually are, however my thought is that at least one of them is tracking spare blocks (which usually starts at 100 and works backwards). Once a drive starts dipping into spare blocks it's usually the end of the line. The original S700/S700 Pro series mimicked the Crucial MX300 in using Micron's 32L/384Gb flash which is technically rated for 1500 P/E but the odd density means the 256GB SKU would have 288GiB of flash (necessary not least because that flash was basically the 32L/256Gb MLC with 3-bit cells). This ends up being ~422TB written not accounting for some write amplification (typically 1.5 for consumer = 280TB as a minimum).

 

[criccio]

Yikes. I wish I knew what I had writing to this drive so frequently. Either way, drive is getting pulled out ASAP. Thank you everyone!

 

[Maxx]

Yikes. I wish I knew what I had writing to this drive so frequently. Either way, drive is getting pulled out ASAP. Thank you everyone!

I would definitely consider it. Usually as soon as I see spare blocks being tapped, I retire the drive - but I can't tell for sure here. The flash on this drive can survive several times that many writes but it's variable depending on write amplification (NAND writes vs. host writes), how many bad blocks were from the factory, environment, many other factors. However, as rated with a general WAF, you are within the range of the flash started to fail.

 

[likeman]

still odd that the SSD it self is not logging uncorrected sector counts

but from general googling "Program Fail Count" of some sort seems to be what the AD is and likely why windows NTFS is able to detect it as a failed LBA sector as the write fails and writes it elsewhere or at least trys to

the drive it self is odd ball as it has no predict self health report smart field or/and available/used spare area smart field

 

[Maxx]

If you ignore raw values and look at actual and threshold you'll see that AD appears to start at 100 with 5 being a warning spot. This almost certainly means it's tracing spare block percentage. Once that value drops from the initial 100 you are living on borrowed time and you WILL get write errors on data from that point forward.

The file system will be tracking logical blocks which are likely 512B in size. The physical blocks of the flash after FTL are likely 24MB on that drive. So a single block going bad is a lot of logical filesystem sectors and the sign of a failing block is from trying to write on it throwing errors. I suppose the drive should be logging these, though.

 

[criccio]

At this point I wish I didn't break the SSD's PCB in a few pieces and trash it and rather, sent it to one of you to play with. LOL

 

[Maxx]

Definitely didn't need to break it, lol, SSDs are relatively easy to securely wipe and recycle. There are ways to pull more information from the drive but either way I think it was in decline.