Me and quite surely a lot of Linoders had problems with kernel oopsi'ng by OOM and swapping errors, even having plenty of free memory. The problems are now back and i'd like to request some assistance on the matter.

I'm using Debian Testing/Unstable (apt-pinned on testing)

Kernel is "Linux brf 3.5.2-linode45 #1 SMP Wed Aug 15 14:10:55 EDT 2012 i686 GNU/Linux"

Since i've instaled tor, it started oops'ng again. checked the max memory/bandwidth configs and everything is tame. Stopped tor and rebooted the machine just to see if it would oops on itself like it did in the past and bingo. here's the trace:

------------[ cut here ]------------
WARNING: at net/core/skbuff.c:3414 skb_try_coalesce+0x293/0x2d0()
Modules linked in:
Pid: 0, comm: swapper/0 Not tainted 3.5.2-linode45 #1
Call Trace:
[<c012d308>] ? warn_slowpath_common+0x78/0xb0
[<c0595fd3>] ? skb_try_coalesce+0x293/0x2d0
[<c0595fd3>] ? skb_try_coalesce+0x293/0x2d0
[<c012d35b>] ? warn_slowpath_null+0x1b/0x20
[<c0595fd3>] ? skb_try_coalesce+0x293/0x2d0
[<c06424bc>] ? tcp_queue_rcv+0xdc/0x130
[<c0645fae>] ? tcp_rcv_established+0x1de/0x5d0
[<c064c83b>] ? tcp_v4_do_rcv+0xbb/0x190
[<c064cfd5>] ? tcp_v4_rcv+0x6c5/0x7b0
[<c062e637>] ? ip_local_deliver_finish+0x97/0x220
[<c062e5a0>] ? ip_rcv+0x330/0x330
[<c062e007>] ? ip_rcv_finish+0xd7/0x340
[<c059d4c3>] ? __netif_receive_skb+0x2c3/0x350
[<c059ed9f>] ? netif_receive_skb+0x1f/0x70
[<c0539b91>] ? handle_incoming_queue+0x1a1/0x270
[<c0539e84>] ? xennet_poll+0x224/0x570
[<c059f6ea>] ? net_rx_action+0xea/0x1a0
[<c01348dc>] ? __do_softirq+0x7c/0x110
[<c0134860>] ? irq_enter+0x70/0x70
<IRQ> [<c01346b6>] ? irq_exit+0x66/0x90
[<c04c50bd>] ? xen_evtchn_do_upcall+0x1d/0x30
[<c072edc7>] ? xen_do_upcall+0x7/0xc
[<c01013a7>] ? hypercall_page+0x3a7/0x1000
[<c010606f>] ? xen_safe_halt+0xf/0x20
[<c011006c>] ? default_idle+0x1c/0x40
[<c011028a>] ? cpu_idle+0x4a/0x80
[<c08e2a9a>] ? start_kernel+0x2e3/0x2e8
[<c08e2632>] ? kernel_init+0x127/0x127
[<c08e5a7c>] ? xen_start_kernel+0x55d/0x565
---[ end trace 6685fabc29338609 ]---

Any lights on this topic?

TY in advance.

Note that it's in the IPv4 stack, which agrees with what other people have seen elsewhere (viewtopic.php?f=20&t=7805).

It's likely to be harmless and just results in a packet retransmission (so may slightly slow down your network connections). But it is alarming

_________________
Rgds
Stephen
(Linux user since kernel version 0.11)

Yeah, i know it's being harmless because my server keeps running but it's not normal to be oopsing everyday. I would like a rock stable operating system i can rely on. Anyone is having these problems still? Did you guys changed kernels and this went away? What kernels are you guys using?

Thanks in advance!

I don't use the linode kernels I use the distribution kernels and encountered the same problem on Ubuntu 12.04. Setting vm.min_free_kbytes=6144 worked on most boxes some I needed to go as high as 8192. I've not see one of these errors in quite a while.

_________________
Paid support
How to ask for help
1. Give details of your problem
2. Post any errors
3. Post relevant logs.
4. Don't hide details i.e. your domain, it just makes things harder
5. Be polite or you'll be eaten by a grue