| 1 |
On Thu, May 23, 2019 at 7:44 AM <mad.scientist.at.large@××××××××.com> wrote: |
| 2 |
|
| 3 |
> I just put an SSD into my hp dl585g7. I used the provided sata SSD drive |
| 4 |
> connector, it's on the same sata driver that runs the optical drive. It's |
| 5 |
> terribly slow to boot etc. Is this just a slow controller designed to keep |
| 6 |
> up with an optical drive and swamped by an SSD? would pulling the optical |
| 7 |
> drive help? I may wind up hooking it to the internal SAS controller with |
| 8 |
> a suitable cable (there are 2 cables to the plugable hard drive area, i |
| 9 |
> could swap one for another cable as I'm only using 2 hard drives. I'd just |
| 10 |
> need to extend the power from the built in ssd connector. Advice? |
| 11 |
> Suggestions?? At this point it's slow to boot like a thumb drive on USB 2 |
| 12 |
> or an optical drive. I have an identical SSD in my desktop and it runs |
| 13 |
> near the rated 500MB/s so I'm confident the SSD is good. |
| 14 |
> |
| 15 |
|
| 16 |
Have you compared hdparm <device> and hdparm -tT <device> outputs between |
| 17 |
the hp and laptop? |
| 18 |
|
| 19 |
On my system with an SATA SSD, its only; |
| 20 |
# hdparm -tT /dev/sda |
| 21 |
|
| 22 |
/dev/sda: |
| 23 |
Timing cached reads: 7820 MB in 2.00 seconds = 3912.89 MB/sec |
| 24 |
Timing buffered disk reads: 172 MB in 3.13 seconds = 55.02 MB/sec |
| 25 |
|
| 26 |
So 3x slower than the spinning disks on the same box for the buffered disk |
| 27 |
reads.... |
| 28 |
# hdparm /dev/sda |
| 29 |
|
| 30 |
/dev/sda: |
| 31 |
multcount = 0 (off) |
| 32 |
IO_support = 1 (32-bit) |
| 33 |
readonly = 0 (off) |
| 34 |
readahead = 256 (on) |
| 35 |
geometry = 30401/255/63, sectors = 488397168, start = 0 |
| 36 |
|
| 37 |
multicount off is weird, lets see what it reports; |
| 38 |
hdparm -i /dev/sda | grep MaxMultSect |
| 39 |
BuffType=unknown, BuffSize=unknown, MaxMultSect=16, MultSect=off |
| 40 |
|
| 41 |
So it should be possible to set it to 16, |
| 42 |
# hdparm -m 16 /dev/sda |
| 43 |
|
| 44 |
/dev/sda: |
| 45 |
setting multcount to 16 |
| 46 |
Use of -m is VERY DANGEROUS. |
| 47 |
Only the old IDE drivers work correctly with -m with kernels up to at least |
| 48 |
2.6.29. |
| 49 |
libata drives may fail and get hung if you set this flag. |
| 50 |
Please supply the --yes-i-know-what-i-am-doing flag if you really want this. |
| 51 |
Program aborted. |
| 52 |
|
| 53 |
Ok i'm not concerned about speed on this box, so i wont bother. |
| 54 |
Interestingly the spinning disks are multicount off on this box too, so |
| 55 |
that setting doesn't explain the SSD's slowness. |
Archives