Setup: DS923 with a DX517 expansion unit which has a SHR-1 volume consisting of 5x 12tb on it. Mixture of WD Red Plus, Ironwolf and Ironwolf Pros
Issue: I/O write errors consistently occurring on drives in the DX517. Have seen 'bad' drives being reported on slots 3-5 however slot 4 seems to be most common. The WD Red Plus are reporting UltraDMA CRC Errors although the Seagates are not (not clear if the Seagates are just not detecting them or what...)
Troubleshooting carried out to date:
- Tried different eSATA cables (Synology and 3rd Party). Issue still present.
- RMA'd DX517 and replaced with new unit. Issue still present.
- RMA'd DS923 and replaced with new unit. Issue still present.
- Swapped out the larger 3rd Party ECC Ram with the Synology ECC ram provided from factory. Issue still present.
- Full destructive testing in Hard Drive Sentinel of each hard drive when flagged 'bad' by DSM. All report as healthy no bad sectors.
- Exchanged hard drives anyway with new when flagged by DSM as 'bad' (expensive endeavour at this point... up to 3 new drives although symptoms remain same), however symptoms continue. Fairly clear at this point that it is not bad drives.
I have considered power quality/interference however I have a DS920 right next to the DS923 with its own DX517 and the DS920 has been flawless without any issues with its DX517 (also 5 drives on SHR-1). Both NAS are on the same UPS.
I have a ticket open with Synology although it has gone quiet on that front, perhaps like me as they are short of ideas. They initially thought it was a bad eSata port on the DS923 but given the DS923 has been exchanged and the issue is still present, that's evidently not the case.
Does anyone have any experience/insight which might shed some light on what the root cause is to the above? It is naturally frustrating have to repair the volume once or twice per week, and I have already experienced one instance where a second drive is reported 'bad' during the repair, leading to a crashed volume.
Thanks in advance :)