March 18, 20188 yr I have a IBM x3850 X5 running 4 E7-8870 and 64 gigs of RAM with 3 terrabytes of internal HD storage. I have installed Unraid version 6.5.0 and have Plex running in Docker. Upon start up it uses about 8 gigs of RAM used then Cached RAM slowly climbs about 1 to 5 gigs every 15 mins, until it maxes out my RAM. I have no idea why. If i stop all containers RAM still runs at capacity. Network usage goes to 0 and disk and CPU usage do the same, but RAM is still maxed. I have Fix Common Errors installed but says everything is fine. Any help would be greatly appreciated. Edited March 18, 20188 yr by Gothix
March 19, 20188 yr What are you doing to determine RAM usage? Are you looking at System Stats? If so, are you talking about the portion of the graph representing "cached"?
March 19, 20188 yr Author This is how i am seeing the RAM usage. The lack of performance is apparent in its inabiliy to run menus or even one direct stream (my laptop does better).
March 19, 20188 yr The RAM usage looks perfectly normal: https://www.linuxatemyram.com Your syslog is full of these Mar 18 15:13:19 Tower root: ACPI group processor / action LNXCPU:07 is not defined which is probably the problem. Haven't seen that before maybe someone else will know.
March 19, 20188 yr The memory utilisation looks completely normal. Linux operating systems generally use spare RAM as cache since it's faster than disk drives. The important part is the Used section in the lighter colour - that's well under 10%, so not a problem. Also see the utilisation reported under the Dashboard tab. Lastly, a basic outline here - https://www.linuxatemyram.com/
March 19, 20188 yr Here is another thread I found: https://lime-technology.com/forums/topic/51307-acpi-spam-in-syslog/
March 19, 20188 yr Here's the FIX I used for this issue: You need to make a copy of /etc/acpi/acpi_handler.sh file, modify it to remove the offending line and then replace it upon every reboot. One time commands to get it setup, copy the existing acpi handler to your flash drive cp /etc/acpi/acpi_handler.sh /boot/config/acpi_handler.sh Now edit /boot/config/acpi_handler.sh with a Unix Friendly editor so it looks like the following, commenting out line 17 #!/bin/sh # Default acpi script that takes an entry for all actions IFS=${IFS}/ set $@ case "$1" in button) case "$2" in power) /sbin/init 0 ;; *) logger "ACPI action $2 is not defined" ;; esac ;; *) # logger "ACPI group $1 / action $2 is not defined" ;; esac Now have this execute during startup, via /boot/config/go file or via user startup scripts: # fix up acpi handler cp /boot/config/acpi_handler.sh /etc/acpi/acpi_handler.sh && chmod 755 /etc/acpi/acpi_handler.sh Edited March 19, 20188 yr by BRiT
March 19, 20188 yr Author 5 hours ago, BRiT said: Here's the FIX I used for this issue: You need to make a copy of /etc/acpi/acpi_handler.sh file, modify it to remove the offending line and then replace it upon every reboot. One time commands to get it setup, copy the existing acpi handler to your flash drive cp /etc/acpi/acpi_handler.sh /boot/config/acpi_handler.sh Now edit /boot/config/acpi_handler.sh with a Unix Friendly editor so it looks like the following, commenting out line 17 #!/bin/sh # Default acpi script that takes an entry for all actions IFS=${IFS}/ set $@ case "$1" in button) case "$2" in power) /sbin/init 0 ;; *) logger "ACPI action $2 is not defined" ;; esac ;; *) # logger "ACPI group $1 / action $2 is not defined" ;; esac Now have this execute during startup, via /boot/config/go file or via user startup scripts: # fix up acpi handler cp /boot/config/acpi_handler.sh /etc/acpi/acpi_handler.sh && chmod 755 /etc/acpi/acpi_handler.sh im sorry but that is so out of my understanding i have no idea what you are saying or where to start. thank you for your help. if you can make it any easer that would be grate.
March 20, 20188 yr On 3/18/2018 at 6:07 PM, BRiT said: Here's the FIX I used for this issue: You think we should just make this change the default? The reason I left it in was so that user was aware of possible h/w issue causing ACPI events to trigger.
March 20, 20188 yr Author yes it would make it more friendly to very new and inexperienced users. Edited March 20, 20188 yr by Gothix
March 21, 20188 yr 19 hours ago, Gothix said: yes it would make it more friendly to very new and inexperienced users. I was concerned of unexplained performance issues if event is triggering all the time but no messages are output.
March 21, 20188 yr In my situation when it does trigger it's for 1 event but like 32 at a time. It might be related to 1 per CORE, but the CPU Numbers don't really match up for 0x00 - 0x20. A quick google for my data seems to be related to CPU speed adjustments via governor? https://bbs.archlinux.org/viewtopic.php?id=134804 Also looks like others posting about it on other forums with Slackware issues -- http://www.jlaforums.com/viewtopic.php?t=342630236 I had this happen at the following times after I updated to 6.5.0 Mar 18 11:43:53 Mar 18 11:43:54 Mar 18 11:53:17 Mar 18 12:02:38 Mar 18 15:11:14 Mar 18 15:41:18 Mar 18 17:20:31 Mar 18 17:47:06 Mar 18 17:51:21 Mar 18 18:13:37 Mar 18 19:24:32 Mar 18 19:24:45 Mar 18 19:25:00 Mar 18 19:25:42 Mar 18 19:26:40 Mar 18 19:27:07 Mar 18 19:27:23 Mar 18 19:28:23 Mar 18 19:28:51 Mar 18 19:29:30 Mar 18 19:39:00 Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:00 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:01 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:02 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:03 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:04 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:05 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:06 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:07 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:08 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:09 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:0a is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:0b is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:0c is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:0d is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:0e is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:0f is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:1e is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:1f is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:20 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:21 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:22 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:23 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:24 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:25 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:26 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:27 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:28 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:29 is not defined Mar 18 11:53:17 TOWER root: ACPI group processor / action LNXCPU:2a is not defined
March 21, 20188 yr Additional details produced from running command to listen for ACPI events and display them. It just happened now after running the listen command for around 45 minutes. System load is 0.09, 0.08, 0.08. I cant seem to stumble upon a definitive list of what event number means at all. # acpi_listen processor LNXCPU:00 00000082 00000000 processor LNXCPU:01 00000082 00000000 processor LNXCPU:02 00000082 00000000 processor LNXCPU:03 00000082 00000000 processor LNXCPU:04 00000082 00000000 processor LNXCPU:05 00000082 00000000 processor LNXCPU:06 00000082 00000000 processor LNXCPU:07 00000082 00000000 processor LNXCPU:08 00000082 00000000 processor LNXCPU:09 00000082 00000000 processor LNXCPU:0a 00000082 00000000 processor LNXCPU:0b 00000082 00000000 processor LNXCPU:0c 00000082 00000000 processor LNXCPU:0d 00000082 00000000 processor LNXCPU:0e 00000082 00000000 processor LNXCPU:0f 00000082 00000000 processor LNXCPU:1e 00000082 00000000 processor LNXCPU:1f 00000082 00000000 processor LNXCPU:20 00000082 00000000 processor LNXCPU:21 00000082 00000000 processor LNXCPU:22 00000082 00000000 processor LNXCPU:23 00000082 00000000 processor LNXCPU:24 00000082 00000000 processor LNXCPU:25 00000082 00000000 processor LNXCPU:26 00000082 00000000 processor LNXCPU:27 00000082 00000000 processor LNXCPU:28 00000082 00000000 processor LNXCPU:29 00000082 00000000 processor LNXCPU:2a 00000082 00000000 processor LNXCPU:2b 00000082 00000000
March 22, 20188 yr 18 hours ago, BRiT said: I cant seem to stumble upon a definitive list of what event number means at all. From include/acpi/video.h in linux source tree: #define ACPI_VIDEO_NOTIFY_CYCLE 0x82 and from drivers/acpi/acpi_video.c: case ACPI_VIDEO_NOTIFY_CYCLE: /* Cycle Display output hotkey pressed. */ a clue?
March 22, 20188 yr Author still no idea whats going on i did a drive change ( without the above fix (couldn't find the file that needed to be edited...)) and did a reboot and now see no errors but its just not running well all...... all i run is plex and after 7 (2 trans-coded 5 direct ) streams it (plex) freaks out and says "Not enough CPU to trans-code" yet im running less then 10% CPU and only 1.3G of ram.......... tower-diagnostics-20180322-1656.zip Edited March 23, 20188 yr by Gothix
March 23, 20188 yr @limetech thanks for looking. I wasn't sure if those lined up or not. Unfortunately it still doesn't seem to map to anything done in reality, at least for my situation. My system is using the onboard video with a Planar LCD connected via VGA style. The monitor is always turned off except for special events. It has no special buttons. The video chip on the motherboard is an AST 2300 (ASPEED AST2300). The motherboard is the ASRock EP2C602 with dual Xeons; picked up from the following thread. https://lime-technology.com/forums/topic/44374-90-xeon-e5-2670-26ghz-8cores-16threads/ The keyboard is connected via USB Bluetooth and is a Logitech Cordless Elite. It sits unused and untouched except for special events. I tried hitting all sorts of keys on the keyboard but all that mapped were of category BUTTON/* or CD/*.
December 9, 20187 yr Hi BRit, Did you finally resolve this dilemma? I have exactly the same hardware and log file output as you and would like to clean this up. Any help would be appreciated.
December 9, 20187 yr 30 minutes ago, Kevlar75 said: Hi BRit, Did you finally resolve this dilemma? I have exactly the same hardware and log file output as you and would like to clean this up. Any help would be appreciated. It's handled through the changes I posted in this thread.
December 9, 20187 yr On 3/19/2018 at 9:07 AM, BRiT said: Here's the FIX I used for this issue: You need to make a copy of /etc/acpi/acpi_handler.sh file, modify it to remove the offending line and then replace it upon every reboot. One time commands to get it setup, copy the existing acpi handler to your flash drive cp /etc/acpi/acpi_handler.sh /boot/config/acpi_handler.sh Now edit /boot/config/acpi_handler.sh with a Unix Friendly editor so it looks like the following, commenting out line 17 #!/bin/sh # Default acpi script that takes an entry for all actions IFS=${IFS}/ set $@ case "$1" in button) case "$2" in power) /sbin/init 0 ;; *) logger "ACPI action $2 is not defined" ;; esac ;; *) # logger "ACPI group $1 / action $2 is not defined" ;; esac Now have this execute during startup, via /boot/config/go file or via user startup scripts: # fix up acpi handler cp /boot/config/acpi_handler.sh /etc/acpi/acpi_handler.sh && chmod 755 /etc/acpi/acpi_handler.sh OK... I'll do this I just assumed because you continued to post queries on the thread that the script route didn't work out. Thanks for clarifying
December 9, 20187 yr OK, Now I have the problem that my system doesn't have an /acpi folder in the /etc directory and when I search for *.sh files there isn't an acpi_handler.sh file in my system???
December 9, 20187 yr 1 hour ago, Kevlar75 said: Now I have the problem that my system doesn't have an /acpi folder in the /etc directory and when I search for *.sh files there isn't an acpi_handler.sh file in my system??? Because you're using Krusader, you're looking at what's in Krusader's folders (its OS, not unRaid's). You've got to use the command line as specified
December 9, 20187 yr Whoops... rookie mistake. Thanks Squid. All done via the console and vim. The logs look good after a restart. Thanks for everyones input 👍 Edited December 9, 20187 yr by Kevlar75 not applicable
Archived
This topic is now archived and is closed to further replies.