Vero 4K crash when using the network with high loads (no high temp)

Help and Support
1

Hello.

I’ve been having random crashes where the vero 4k+ just shutsdown. It looks like it is when there is high load in the network stack. Sometimes is transmision, sometimes is just sending to the vero through ssh (scp) a big file, or accesing via nfs…

The temperature never goes higher than 60 degrees so is not a thermal shutdown.

I have a ssd disk atached to the vero in the usb port without a hub or any external power. So it could be also related to the power drawn. But kernel messages suggest otherwise.

ene 21 21:47:23 osmcsala kernel: Call trace:
ene 21 21:47:23 osmcsala kernel: Exception stack(0xffffffc074b44740 to 0xffffffc074b44870)
ene 21 21:47:23 osmcsala kernel: 4740: ffffffc04ff53c00 0000007fffffffff ffffffc074b4f550 ffffff800912c4dc
ene 21 21:47:23 osmcsala kernel: 4760: ffffffc053b9cf00 ffffff800a814380 ffffffc050898000 ffffff800a814380
ene 21 21:47:23 osmcsala kernel: 4780: 0000000000000000 ffffff800a814380 8000000000000004 ffffff8009938f02
ene 21 21:47:23 osmcsala kernel: 47a0: 0000000000000000 ffffffc057348000 ffffffc074b447d0 ffffff80099e0538
ene 21 21:47:23 osmcsala kernel: 47c0: ffffffc057348000 ffffff80099e0454 ffffffc074b44860 ffffff80099a9dec
ene 21 21:47:23 osmcsala kernel: 47e0: 0000000000000000 ffffff800a83c000 0000000000000001 0000000000000000
ene 21 21:47:23 osmcsala kernel: 4800: 000000000000000c ffffffc074b44a9f ffffffc04fce0ed8 0000000000000000
ene 21 21:47:23 osmcsala kernel: 4820: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
ene 21 21:47:23 osmcsala kernel: 4840: 0000000000000000 0000000000000000 0000000000000000 0000000095468077
ene 21 21:47:23 osmcsala kernel: 4860: 0000000000000010 00000000cc30f388
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44910+  32][<ffffff800912c4dc>] cgroup_get+0x34/0xe8
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44930+  32][<ffffff8009132f4c>] cgroup_sk_alloc+0x48/0x98
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44950+  48][<ffffff8009934ef4>] sk_clone_lock+0x148/0x254
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44980+  48][<ffffff80099b79a8>] inet_csk_clone_lock+0x30/0xcc
ene 21 21:47:23 osmcsala kernel: [ffffffc074b449b0+  64][<ffffff80099d09bc>] tcp_create_openreq_child+0x3c/0x348
ene 21 21:47:23 osmcsala kernel: [ffffffc074b449f0+  80][<ffffff80099cdb34>] tcp_v4_syn_recv_sock+0xf0/0x2b4
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44a40+ 144][<ffffff80099d11d0>] tcp_check_req+0x35c/0x3dc
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44ad0+ 112][<ffffff80099d0154>] tcp_v4_rcv+0x504/0xb0c
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44b40+  48][<ffffff80099ac370>] ip_local_deliver_finish+0x150/0x214
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44b70+ 112][<ffffff80099ac908>] ip_local_deliver+0x88/0xe8
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44be0+  64][<ffffff80099ac5d0>] ip_rcv_finish+0x19c/0x32c
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44c20+ 128][<ffffff80099acc98>] ip_rcv+0x330/0x394
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44ca0+ 176][<ffffff800994a0b4>] __netif_receive_skb_core+0x634/0x810
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44d50+  48][<ffffff800994aa48>] __netif_receive_skb+0x78/0x7c
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44d80+  48][<ffffff800994aaf4>] netif_receive_skb_internal+0xa8/0xb4
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44db0+  48][<ffffff800994b4ec>] napi_gro_receive+0x10c/0x140
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44de0+ 128][<ffffff800952b93c>] stmmac_poll+0x5d8/0x824
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44e60+ 176][<ffffff800994bcb4>] net_rx_action+0x128/0x33c
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44f10+ 144][<ffffff8009081960>] __do_softirq+0x250/0x35c
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44fa0+  32][<ffffff80090a579c>] irq_exit+0x90/0xdc
ene 21 21:47:23 osmcsala kernel: [ffffffc074b44fc0+  64][<ffffff80090f62fc>] __handle_domain_irq+0x88/0xbc
ene 21 21:47:23 osmcsala kernel: [ffffffc074b45000+  64][<ffffff8009081568>] gic_handle_irq+0x68/0xa4
ene 21 21:47:23 osmcsala kernel: Exception stack(0xffffff80203a37e0 to 0xffffff80203a3910)
ene 21 21:47:23 osmcsala kernel: 37e0: 0000000000009b70 0000000000000064 0000000000000000 0000000000000000
ene 21 21:47:23 osmcsala kernel: 3800: 0000000060000000 000000001fb0f810 0000000000000c98 0000000000000278
ene 21 21:47:23 osmcsala kernel: 3820: ffffffc053a2ae80 ffffff80203a37f0 00000000000008a0 0000000000000024
ene 21 21:47:23 osmcsala kernel: 3840: 000000000000009a 0000000000000000 0000000000000000 0000000000000000
ene 21 21:47:23 osmcsala kernel: 3860: ffffff80091fd49c 0000000000000000 0000000000000000 000000001fb417b4
ene 21 21:47:23 osmcsala kernel: 3880: 000000000000bb80 0000000000000000 0000000000036ee8 ffffff800a7a50c0
ene 21 21:47:23 osmcsala kernel: 38a0: ffffff800a92a000 ffffff8009ef18e9 ffffff8009e70fb4 ffffff80203a3b40
ene 21 21:47:23 osmcsala kernel: 38c0: 0000000000000000 ffffff80203a3910 ffffff80093cf7c0 ffffff80203a3910
ene 21 21:47:23 osmcsala kernel: 38e0: ffffff80093cf7ec 0000000080000145 0000000000000000 0000000000000000
ene 21 21:47:23 osmcsala kernel: 3900: 0000007fffffffff 0000000000000000

Here are the full logs:
https://paste.osmc.tv/bihatarisu

Let me know if you need aditional data.
Thanks in advance.

2

Do you think, if the Vero has a moment of power instability, network wouldn’t be effected??

First step try powering the drive, even though it’s SSD and “shouldn’t” need power. There is no way to really “disprove” if the power used by the drive, combine with other factors could effect the NIC. Vero has USB2, ergo lower limit of amps. Most external harddrives are built for USB3, where there is in the standard a higher limit for amps delivered.

What filesystem is on that drive? Depending on FS, there is different CPU loads for writing and that goes for network transfer protocol used as well. Which means CPU will consume different amounts of power, while having a external drive powered by Vero4k(+) at the same time can spell disaster.

3

It’s ext4
I don’t have rigth now a powered hub but i’ll make some tests writing directly to the internal ssd with transmision or via ssh and see if still happens. And as soon as I can I will give it a shot to a powered hub.
Thanks

4

This will be your problem.

5

Yes. That was the problem. Looks like vero 4k+ can’t handle writing to a ssd without external power with high usage.