This forum uses cookies
This forum makes use of cookies to store your login information if you are registered, and your last visit if you are not. Cookies are small text documents stored on your computer; the cookies set by this forum can only be used on this website and pose no security risk. Cookies on this forum also track the specific topics you have read and when you last read them. Please confirm that you accept these cookies being set.

Spacelynk Cpu load high
#1
Hello,
I have a very high cpu load on customer Spacelynk. The Spacelynk wasn't installed by me, but I think it is very strange ! I tried to do a Factory reset without difference ! Do you have Idea ? I also upgrade to 2.1.1 firmware without any difference !
There is no script.. If I put the backup on my test Spacelynk th load is 0,1%

What is fun, is that when rebooting, it start from 0.5 then it's growing in the 3 min to more than 1

CPU/IO: 1.52 0.96 0.54, Memory: 11%, KNX/TP: OK


Memory looks good:
Total system memory
122.7M
Used
13.0M (11%)
Free
109.6M (89%)
Buffered
1.2M
Cached
16.1M

system log

Code:
May 28 08:34:25 spaceLYnk user.info sysinit: Generating selfsigned certificate with subject 'CN=spaceLYnk;' and validity 2018-05-28 08:34:00-2033-05-18 03:33:20
May 28 08:34:00 spaceLYnk user.info sysinit: Generating RSA private key, 2048 bit long modulus
May 28 08:33:59 spaceLYnk user.info sysinit: Generating selfsigned certificate with subject 'CN=spaceLYnk;' and validity 2018-05-28 08:33:22-2033-05-18 03:33:20
May 28 08:33:30 spaceLYnk kern.notice kernel: [ 25.621675] random: nonblocking pool is initialized
May 28 08:33:22 spaceLYnk user.info sysinit: Generating RSA private key, 2048 bit long modulus
May 28 08:33:19 spaceLYnk user.info sysinit: Configuring touch.
May 28 08:33:19 spaceLYnk user.info sysinit: Configuring fbeditor20.
May 28 08:33:14 spaceLYnk user.info sysinit: Installing touch (20170810) to root...
May 28 08:33:12 spaceLYnk user.info sysinit: Installing fbeditor20 (20170810) to root...
May 28 08:33:11 spaceLYnk kern.info kernel: [ 6.443255] fec 800f0000.ethernet eth0: Link is Up - 100Mbps/Full - flow control off
May 28 08:33:09 spaceLYnk user.info sysinit: uci: Entry not found
May 28 08:33:09 spaceLYnk user.info sysinit: uci: Entry not found
May 28 08:33:09 spaceLYnk kern.info kernel: [ 4.749319] EXT4-fs (mmcblk0p3): mounted filesystem with ordered data mode. Opts: nodelalloc
May 28 08:33:09 spaceLYnk user.info sysinit: /dev/mmcblk0p3: clean, 14/131072 files, 25391/524288 blocks
May 28 08:33:09 spaceLYnk user.info sysinit: e2fsck 1.43.7 (16-Oct-2017)
May 28 08:33:09 spaceLYnk kern.info kernel: [ 4.447035] fec 800f0000.ethernet eth0: Freescale FEC PHY driver [SMSC LAN8710/LAN8720] (mii_bus:phy_addr=800f0000.ethernet-1:00, irq=-1)
May 28 08:33:07 spaceLYnk kern.info kernel: [ 3.178512] fec 800f0000.ethernet eth0: Freescale FEC PHY driver [SMSC LAN8710/LAN8720] (mii_bus:phy_addr=800f0000.ethernet-1:00, irq=-1)
May 28 08:33:07 spaceLYnk kern.info kernel: [ 2.833436] watchdog timeout 60 seconds
May 28 08:33:07 spaceLYnk kern.info kernel: [ 1.698404] EXT4-fs (mmcblk0p2): re-mounted. Opts: nodelalloc
May 28 08:33:07 spaceLYnk kern.info kernel: [ 0.898524] Freeing unused kernel memory: 140K
May 28 08:33:07 spaceLYnk kern.info kernel: [ 0.897745] VFS: Mounted root (ext3 filesystem) readonly on device 179:2.
May 28 08:33:07 spaceLYnk kern.info kernel: [ 0.897569] EXT4-fs (mmcblk0p2): mounted filesystem with ordered data mode. Opts: (null)
May 28 08:33:07 spaceLYnk kern.info kernel: [ 0.767230] mmcblk0: p1 p2 p3
May 28 08:33:07 spaceLYnk kern.info kernel: [ 0.764655] mmcblk0: mmc0:59b4 00000 3.67 GiB
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.762780] mmc0: new high speed SDHC card at address 59b4
May 28 08:33:06 spaceLYnk kern.warn kernel: [ 0.760229] mmc0: host does not support reading read-only switch, assuming write-enable
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.753740] Waiting for root device /dev/mmcblk0p2...
May 28 08:33:06 spaceLYnk kern.err kernel: [ 0.753577] Warning: unable to open an initial console.
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.751642] rtc-ds1307 0-006f: setting system clock to 2018-05-28 08:33:05 UTC (1527496385)
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.745389] NET: Registered protocol family 17
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.722434] mxs-mmc 80010000.ssp: initialized
May 28 08:33:06 spaceLYnk kern.warn kernel: [ 0.690655] 80010000.ssp supply vmmc not found, using dummy regulator
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.689497] usbcore: registered new interface driver btusb
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.683981] lirc_dev: IR Remote Control driver registered, major 248
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.682643] i2c /dev entries driver
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.681180] rtc-ds1307 0-006f: 64 bytes nvram
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.681059] rtc-ds1307 0-006f: rtc core: registered mcp7941x as rtc0
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.677569] g_ether gadget: g_ether ready
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.677256] g_ether gadget: Ethernet Gadget, version: Memorial Day 2008
May 28 08:33:06 spaceLYnk kern.warn kernel: [ 0.677002] using random host ethernet address
May 28 08:33:06 spaceLYnk kern.warn kernel: [ 0.676925] using random self ethernet address
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.676753] usb0: MAC c6:ee:90:93:67:f8
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.676444] usb0: HOST MAC 46:bb:37:f9:92:13
May 28 08:33:06 spaceLYnk kern.warn kernel: [ 0.673143] using random host ethernet address
May 28 08:33:06 spaceLYnk kern.warn kernel: [ 0.673053] using random self ethernet address
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.670904] hub 1-0:1.0: 1 port detected
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.670668] hub 1-0:1.0: USB hub found
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.668520] usb usb1: SerialNumber: ci_hdrc.0
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.668490] usb usb1: Manufacturer: Linux 4.4.109 ehci_hcd
May 28 08:33:06 spaceLYnk kern.info kernel: [ 0.668459] usb usb1: Product: EHCI Host Controller
cat /proc/meminfo

Code:
* table:
[1]
 * string: MemTotal:         125624 kB
[2]
 * string: MemFree:           93332 kB
[3]
 * string: MemAvailable:     108536 kB
[4]
 * string: Buffers:            1340 kB
[5]
 * string: Cached:            16880 kB
[6]
 * string: SwapCached:            0 kB
[7]
 * string: Active:            19408 kB
[8]
 * string: Inactive:           4400 kB
[9]
 * string: Active(anon):       5820 kB
[10]
 * string: Inactive(anon):       68 kB
[11]
 * string: Active(file):      13588 kB
[12]
 * string: Inactive(file):     4332 kB
[13]
 * string: Unevictable:           0 kB
[14]
 * string: Mlocked:               0 kB
[15]
 * string: SwapTotal:             0 kB
[16]
 * string: SwapFree:              0 kB
[17]
 * string: Dirty:                16 kB
[18]
 * string: Writeback:             0 kB
[19]
 * string: AnonPages:          5596 kB
[20]
 * string: Mapped:             3660 kB
[21]
 * string: Shmem:               308 kB
[22]
 * string: Slab:               4628 kB
[23]
 * string: SReclaimable:       1640 kB
[24]
 * string: SUnreclaim:         2988 kB
[25]
 * string: KernelStack:         352 kB
[26]
 * string: PageTables:          248 kB
[27]
 * string: NFS_Unstable:          0 kB
[28]
 * string: Bounce:                0 kB
[29]
 * string: WritebackTmp:          0 kB
[30]
 * string: CommitLimit:       62812 kB
[31]
 * string: Committed_AS:       9496 kB
[32]
 * string: VmallocTotal:     901120 kB
[33]
 * string: VmallocUsed:           0 kB
[34]
 * string: VmallocChunk:          0 kB

Could it be the number of device connected ? There is 4 Touch HP computer I think !
-----------
FRANCE SMARTHOME & SMARTBUILDING INTEGRATION
SE ECO EXPERT
Reply
#2
Start with disconnecting client devices and checking KNX/IP bus load.
Reply
#3
(28.05.2018, 09:44)admin Wrote: Start with disconnecting client devices and checking KNX/IP bus load.

knx bus load is 10% average. Never above 20% (except when programming with ets).
I disconnect the Touch screen HP slate 21 (android), same cpuload that increase from 0.3 at boot to 1 after 5 minutes.

Ask my self about hardware issue or something else !

I did factory restore from web interface. Should I do something else ?

PS: For info, I know there is 2 knx Ip device on the same network (2 other line as backbone router). I tried to disable Ip stuff from Spacelynk without any difference
-----------
FRANCE SMARTHOME & SMARTBUILDING INTEGRATION
SE ECO EXPERT
Reply
#4
Are you sure there are no scripts running at all? Infinite loop in/between scripts or a resident script with sleep time set to 0 can consume whole CPU.
Reply
#5
looks like there is multiple send of telegram (with routing number wich increase). For example a telegram from a temperature sensor is send on the bus. But when i log, i see it more than 5 time.
The line 1 is with Schneider Ip router
The line 2 is with Schneider Ip router

The Spacelynk is on line1, but also with IP interface (multicast adress are the same than the 2 Ip router). I will try to put a different multicast Ip adress to test.
-----------
FRANCE SMARTHOME & SMARTBUILDING INTEGRATION
SE ECO EXPERT
Reply
#6
Hi
It looks to me that you have KNX loop. Disable KNX IP features.
BR
------------------------------
Ctrl+F5
Reply
#7
The Ip feature is needed to program Product when away from the customer building (I use script to enable/disable).

Looks like it's working when changing multicast adress. The standar one is 224.0.23.12. So i imagine that the Spacelynk see it on the Tp part, and also on the multicast Ip adress, then maybe try to send it again on Multicast...

I change it on spacelynk:
224.0.23.13

I let the both Ip router on 224.0.23.12.

Now I have telegram send only one on the bus monitor and i know can program with Spacelynk (I couldn't before, timeout error).

The second option I had what to disable Tp feature, and only use Ip with both router.

What do you think about this study case
-----------
FRANCE SMARTHOME & SMARTBUILDING INTEGRATION
SE ECO EXPERT
Reply
#8
Hi

What for do you have KNX IP router and SL connected to the same line?
Different multicast is another way of solving this design issue Wink

BR
------------------------------
Ctrl+F5
Reply
#9
yes, the design is very strange.

They have 2 building and instead of using tp coupler they used 2 Ip router.

What you said is to switch one Ip router and using Spacelynk as router? I thing it's less failsafe to use Linux software instead of Knx tp product to do the job.
-----------
FRANCE SMARTHOME & SMARTBUILDING INTEGRATION
SE ECO EXPERT
Reply
#10
Yes, but it is your choice.
If you keep both just take this in to account that 2 routers ether must have different multicast or KNX IP disabled otherwise KNX loop is created.
------------------------------
Ctrl+F5
Reply
#11
Shut down the filter tables in the SL to break the loop. :-)
Reply
#12
(04.12.2018, 13:38)equalizer Wrote: Shut down the filter tables in the SL to break the loop.  :-)

I disable multicast on SL Rolleyes
-----------
FRANCE SMARTHOME & SMARTBUILDING INTEGRATION
SE ECO EXPERT
Reply


Forum Jump: