Intel 530 SSD very slow (about 2 Mbyte/s) read and write only in a HP Z230

[BGábo]

Hello dear readers,

I have an issue with an Intel 530 120 GiB SSD. We use it since some years as a database drive in a 24/7 working, MSSQL server on Windows 7. The drive is a second one in a HP Z230 series tower, with gen 4 Intel i5 CPU and corresponding hardware. Around a year before, we experienced ridiculous slowdown during log file deletion. After deleting more than 10 files in a row, the SSD slowed down as seriously as MSSQL server hanged up for a minute. What i have tried already:

-Clean out the PC, clean RAM contacts, fans, redo all cable conenctions and change cables

-Test RAM

-Updated system UEFI

-Let Intel SSD Optimizer do its job-no new firmware, no SMART error, TRIM works fine, SSD Optimization is done

-Changed SATA mode from RAID to AHCI and for further tests, compatible IDE

-Run off from Hiren's Boot CD Windows XP and did a disk image copy on the first HDD to overcome the situation. The speed was extremly slow, it took 5 hours to copy 40 GB of disk image.

After i took the drive in an USB3 enclosure on a differenct PC and everything seems fine, no slowdown, and speed is normal.

So, all in one, i have to figure out if the SSD is defective, or facing with a firmware related problem, or maybe the server has something wrong with the SATA controller (not really think-as the other HDD works fine)

[idata]

Hello didyman,

Thank you for your reply.Based on the log files you provided we have detected that your Intel® SSD 530 Series already has the latest version of the firmware and is being exposed to a workload a lot higher than the one that this drive was designed for, this being the cause of the media wear out and the behavior that this device is presenting.As previously mentioned your Intel® SSD 530 Series is qualified as a "Consumer SSD" and was not designed to be used on a server/data-center environment, and using your drives in an out of specs environment void your warranty. To explain the reason why the misuse would void the warranty; when a customer uses a consumer drive in a data center usage, is exposing the drive to different workloads that involve different endurance and therefore surpassing the specs. Our consumer drives are expected for client/ basic-gaming usage. That is why we offer data center drives since those have specs which have been validated for enterprise usage.We hope you find this information useful.Best regards,Josh B.Intel Customer Support TechnicianUnder Contract to Intel Corporation

[BGábo]

Hello, Josh,

Thank You for your answer. I would like to sum up my thoughts:

-drive is not under warranty. Any warranty-related thing is without meaning.

-drive has around 74 % wearout . Let's be true: it's not that bad for a 5 years old SSD in this task.

But what makes this conversation a bit out of the line: the mentioned erroneous behaviour is only expreinced in one, and in that only one PC. If what You say is truly the cause of the error, the phenomenon SHOULD travel to PC from PC, i would have to experience the same slowdown in any other configuration. But the SSD works since in the test PC, with the same filesystem, with the same database on it, works in USB rack and if i put it back in the server, it fails. Please, tell me, how it's connected to wear level?

Sorry for being a bit frustrated. I'm really appreciate time and effort You put in your answer, but we have sticked to that "this-is-not-an-enterprise-SSD" statement, what i agree, but the current situation cannot proove anything regarding it, except wear level, wich is IMHO bounded to the fact the SSD contains more percentage of data has to be trimmed because of small, but frequent writes. Bump.

[idata]

Hello didyman,

Thank you for your reply.

As mentioned on our previous interaction based on the log files you provided we have detected that your Intel® SSD 530 Series is exposed to a workload a lot higher than the one that this drive was designed for, you can check this on the SMART log information by looking the"Total NAND Writes" and the "Host Writes" this information can give you an idea of the workload that this drive is being exposed, another point that you can check is the "End-to-End Error Detection Count" that is lower than the "Normalized" that should be "100" this is a signal of degradation on the drive (media wear out) and this parameter may indicate a future drive failure.

Such end-to-end errors usually caused by one of the following situations:

1) The hard disk internal cache memory is damaged (media wear out) which may result in data corruption and other issues.

2) There is a problem with the connection line between the disk and the computer. Sometimes the issue is related to cable/connection issue and this can cause performance degradations and may cause data corruption.

3) There is a problem on the side of the computer (overclocking, RAM issues, chipset/USB controller (which used to control the hard disk) overheat, etc.)

If the drives used as internal ones, it is possible that the computer shows "S.M.A.R.T. failure" upon starting and also Windows may show hard disk problem in such situation.

Please take into consideration that these drives are not tested or validated to be used as USB storage.

As previously mentioned since your drives are being used on an environment out of the specs (data center environment or USB) we cannot predict how this device will behave could be better or worst we do not have any test done to determine this.

At the moment that you move this drive to a new system besides exposing this HDD to a different workload environment it can solve the issue if it is related to the points 2 and 3.

Since your issue only occurs on a server environment and not in any other consumer PC environment we cannot provide you with any other recommendation that the ones we already shared.

Best regards,

Josh B.

Intel Customer Support TechnicianUnder Contract to Intel Corporation

[BGábo]

Maybe i wasn't detailed enough in my first post. The issue occurs not only in server environment.

The tests i made with the PC in question, are all made booted from Hirens boot CD, in that case, the database server wasn't run, we can't talk about server environment or workload, that was the very fundamental meaning of booting from Hirens. With the very same manner, i tested the SSD in other PC and finally, test run under Windows, with stopped MSSQL services and for another run, in an USB enclosure (supported or not-it worked) without the symptom with my PC. The SSD now is in another (test-) server, supporting the same database as in the original, with simulated workload similar to that one. I think, i really made a cross-test to determine circumstances.

Maybe i forgot to mention enough detailed, there is absolutely no SMART error, nor in UEFI, nor in Intel SSD Toolbox, nor in Event Viewer. By the way, that bigger part of your reply is exactly found in a Hard Disk Sentinel article from 2016. 🙂

NAND writes are high, more than 80 TB (this maybe shows the result of much inside operations because of much more small writes and TRIM), but host writes are only 5 TB. My home rig's Samsung 840 pro SSD has 18 TB of host writes (and i'm not a torrent guy, this is a net and sometimes game PC).

End to end detection really shows something, but as far as it is not concerned as an error in SMART status, i wonder, if that level could be a reason behind this selectively.

This is a factory set server and widely tested, reliability focused environment (except the SSD 🙂 ). Any kind of overclocking, altering of key parameters are strictly forbidden. Cables are of highest quality shielded ones and as i pointed out in my first post, already tried to replace them, and also on different SATA ports.

[idata]

Hello didyman,

Thank you for your reply.

As mentioned on our previous interaction based on the log files you provided we have detected that your Intel® SSD 530 Series is exposed to a workload a lot higher than the one that this drive was designed for, the "End-to-End Error Detection Count" is showing a parameter lower than the "Normalized" that should be "100" this is a signal of degradation on the drive (media wear out) and this may indicate a future drive failure.

Our last recommendation in your case since your drive has more than 5 years being exposed to an out of specs environment plus that you are experiencing various issues on different environment and since the SMART logs provided by you are showing different errors it is to replace the drive in order to avoid data loss and further problems.

Best regards,

Josh B.

Intel Customer Support TechnicianUnder Contract to Intel Corporation