| View previous topic :: View next topic |
| Author |
Message |
Anon-E-moose
Watchman
Joined: 23 May 2008
Posts: 6098
Location: Dallas area
|
 Posted: Fri Jun 09, 2023 3:53 pm Post subject: |
 |
|
| Black wrote: | | Anon-E-moose wrote: | | Do you have nvme-cli installed? It has lots of useful options for nvme investigation. |
Yes I do, but I haven't used that before. Any hint as to which commands to look at?
Thank you (and everyone else)! |
There are a few *log commands, get-log, error-log, etc, get* commands are usually safe 
_________________
PRIME x570-pro, 3700x, 6.1 zen kernel
gcc 13, profile 17.0 (custom bare multilib), openrc, wayland |
|
| Back to top |
|
 |
Black
Apprentice
Joined: 10 Dec 2002
Posts: 158
Location: Québec, Canada
|
| Posted: Sun Jun 11, 2023 2:24 pm Post subject: |
|
|
I have put /tmp on tmpfs and set the kernel parameter | Code: | | nvme_core.default_ps_max_latency_us=2000 |
, yet the Data Units Written (divided by the number of hours between 2 snapshots) for the past 24 hours have almost doubled.
| eccerr0r wrote: | Searching the web, I get a lot of hits on kingston ssds having this behavior...
Currently I only have Intel, Samsung(mPCIe) Patriot (mPCIe), HP, and Micron/Crucial SSDs ... they don't seem to exhibit this behavior though the Samsung I accidentally swap stormed on and ate a chunk of its life ... |
At this point it appears very likely that this is an issue with Kingston. I'm ordering a Crucial NVMe and I'll report back once I have a bit of data.
I'll still try to run those nvme-cli commands and see what it says. |
|
| Back to top |
|
|
Black
Apprentice
Joined: 10 Dec 2002
Posts: 158
Location: Québec, Canada
|
| Posted: Sat Jun 24, 2023 12:09 pm Post subject: |
|
|
It's been a few weeks, and while I have the new nvme on my desk, I haven't had the time to put it in.
That said, I have put /tmp on tmpfs a while ago, and the average data written had dropped from ~60,000KB/hour to around 5KB/hour. I did find out that Google Chrome likes to write a lot in /tmp, even when it's not doing anything (such as just being open while I sleep). I have also added the | Code: | | --disk-cache-dir="/tmp/black/google-chrome" |
command-line parameter to the Chrome shortcut, though it should not matter since /home is not on the NVMe to begin with. (I also added a cronjob to clear that cache every hour because it grows fast.)
As for the nvme-cli commands, I haven't been able to find any new information.
nvme error-log /dev/nvme0 is full is identical entries:
| Code: | Entry[63]
.................
error_count : 0
sqid : 0
cmdid : 0
status_field : 0(Successful Completion: The command completed without error)
phase_tag : 0
parm_err_loc : 0
lba : 0
nsid : 0
vs : 0
trtype : The transport type is not indicated or the error is not transport related.
csi : 0
opcode : 0
cs : 0
trtype_spec_info: 0
log_page_version: 0
................. |
get-log is no help:
| Code: | blackphoenix ~ # nvme get-log -l 512 /dev/nvme0
NVMe status: Invalid Log Page: The log page indicated is invalid(0x109) |
smart-log shows the same information as smartctl:
| Code: | Smart Log for NVME device:nvme0 namespace-id:ffffffff
critical_warning : 0x4
temperature : 31 °C (304 K)
available_spare : 100%
available_spare_threshold : 10%
percentage_used : 100%
endurance group critical warning summary: 0
Data Units Read : 320,040 (163.86 GB)
Data Units Written : 2,170,006,961 (1.11 PB)
host_read_commands : 11,862,159
host_write_commands : 8,589,274,821
controller_busy_time : 97,527
power_cycles : 48
power_on_hours : 17,812
unsafe_shutdowns : 18
media_errors : 0
num_err_log_entries : 0
Warning Temperature Time : 0
Critical Composite Temperature Time : 0
Thermal Management T1 Trans Count : 0
Thermal Management T2 Trans Count : 0
Thermal Management T1 Total Time : 0
Thermal Management T2 Total Time : 0 |
|
|
| Back to top |
|
|
eccerr0r
Watchman
Joined: 01 Jul 2004
Posts: 9679
Location: almost Mile High in the USA
|
| Posted: Fri Jul 28, 2023 5:36 pm Post subject: |
|
|
holy crap so it was chrome that was writing that much crap to /tmp all this time, 60MB/hour?!? ugh... shared memory file? hmm.
Yeah have to use tmpfs a lot to placate software that thinks it should be able to write and rewrite (state files for crash mitigation)?
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching? |
|
| Back to top |
|
|
pietinger
Moderator
Joined: 17 Oct 2006
Posts: 4148
Location: Bavaria
|
| Posted: Fri Jul 28, 2023 5:47 pm Post subject: |
|
|
It is not only the browser ... Therefore I am using this solution: https://wiki.gentoo.org/wiki/SSD#XDG_cache_on_tmpfs
Now my ~/.cache is really empty; it is all now in /tmp/peter ... of course a tmpfs:
| Code: | /tmp/peter/.cache # ls -al
insgesamt 27576
drwxr-xr-x 13 peter users 340 28. Jul 19:02 .
drwxr-xr-x 3 peter users 60 28. Jul 13:03 ..
drwxr-xr-x 2 peter users 60 28. Jul 13:03 akonadi_ical_resource_0
drwxr-xr-x 3 peter users 60 28. Jul 13:06 falkon
drwxr-xr-x 2 peter users 300 28. Jul 19:43 favicons
-rw-r--r-- 1 peter users 10547304 28. Jul 17:08 icon-cache.kcache
drwxr-xr-x 2 peter users 160 28. Jul 19:39 kio_http
drwx------ 3 peter users 60 28. Jul 13:05 konqueror
drwxr-xr-x 3 peter users 60 28. Jul 13:03 ksplash
-rw-r--r-- 1 peter users 755290 28. Jul 13:03 'ksycoca5_de_9FMAPGKp1wKkh1j4orZoPhe4zb4='
drwxr-xr-x 3 peter users 60 28. Jul 15:21 kwin
drwxr-xr-x 110 peter users 2220 28. Jul 19:28 mesa_shader_cache
drwxr-xr-x 3 peter users 60 28. Jul 13:03 plasmashell
-rw------- 1 peter users 46301 28. Jul 19:02 plasma-svgelements
drwxr-xr-x 3 peter users 60 28. Jul 14:07 plasma-systemmonitor
-rw-r--r-- 1 peter users 16875624 28. Jul 19:28 plasma_theme_default_v5.108.0.kcache
drwxr-xr-x 2 peter users 440 28. Jul 15:21 qtshadercache-x86_64-little_endian-lp64 |
|
|
| Back to top |
|
|
Goverp
Advocate
Joined: 07 Mar 2007
Posts: 2007
|
| Posted: Sat Jul 29, 2023 8:44 am Post subject: |
|
|
Interesting. My .cache has entries back to 2018, and is 1.4 GB. Somehow I suspect there's a lot of cruft there. I may try wiping the lot...
_________________
Greybeard |
|
| Back to top |
|
|
eccerr0r
Watchman
Joined: 01 Jul 2004
Posts: 9679
Location: almost Mile High in the USA
|
| Posted: Sat Jul 29, 2023 1:57 pm Post subject: |
|
|
I wouldn't indiscriminately wipe .cache, seems it also holds temp files that were downloaded (and could be downloaded again) or precomputed files (and could be precomputed again) but kept for speed.
That being said my ~/.cache is 5GB (!) alas I don't see the SSD chomping... I don't use Chrome...
I also wonder what's XDG_CACHE_HOME compared to just putting cache files in $XDG_RUNTIME_DIR which should be a tmpfs on /run ? Could there be something in XDG_CACHE_HOME that should be kept from boot to boot for speed purposes?
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching? |
|
| Back to top |
|
|
Goverp
Advocate
Joined: 07 Mar 2007
Posts: 2007
|
| Posted: Sun Jul 30, 2023 6:55 am Post subject: |
|
|
| eccerr0r wrote: | | I wouldn't indiscriminately wipe .cache, seems it also holds temp files that were downloaded (and could be downloaded again) or precomputed files (and could be precomputed again) but kept for speed.... |
And that's the usual trouble with persistent cache - it may be fast, but it may be old and out of date.
I've moved my .cache to old.cache. So far nothing's broken. I need to try a few more critical apps - I experienced various Android apps losing their memory when I did similar to my (old) tablet computer, which of course is App writers not understanding "cache".
_________________
Greybeard |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Other Things Gentoo
