On 4/2/25 17:21, Naveen Kumar P wrote: > On Wed, Apr 2, 2025 at 1:29 PM Damien Le Moal <dlemoal@xxxxxxxxxx> wrote: >> >> On 4/2/25 15:40, Naveen Kumar P wrote: >>> Questions: >>> 1. What could cause periodic WRITE DMA timeouts followed by link resets? >> >> The drive not responding (drive failing for whatever reasons) >> >>> 2. Could this be a hardware-related issue (e.g., cabling, drive aging) >>> or a kernel bug? >> >> It always can be a bug, but it looks like this is with a kernel version 4.4, >> which is really old. Please try with the latest kernel. We do not debug older >> kernels. >> >> Hardware (SSD) failing is more likely though. Your SSD is being operated with >> NCQ turned off: > Thank you for your response. Would re-enabling NCQ (if possible) help > improve performance, or would it likely introduce instability? NCQ was disabled for this SSD + adapter because it was deemed unstable. So re-enabling it (removing the quirk) will likely make things worse. > I observed that when the system is in this state (frequent WRITE DMA > timeouts and link resets), running aplay from alsa-utils debian > package results in the following error: > aplay: pcm_write:2086: write error: Input/output error > > Could these storage timeouts be affecting aplay? Is there a known > relationship between these two failures? This sounds like a hardware problem. Maybe your power supply is not adequate or unstable ? In any case, unless you retest with a recent kernel, nothing much we can do. -- Damien Le Moal Western Digital Research
