[ovs-discuss] OVS - PDUMP: Pdump initialization failure in different container

Rajesh Kumar rajesh.kumar at certesnetworks.com
Fri Sep 27 06:49:57 UTC 2019


Hi Ian,


Thanks for the reply,


We are sharing hugepages between ovs and pdump pods using pod deployment files.

The following is the pdump deployment yaml file where access to hugepages (/dev/hugepages-cf) is specified, similarly we do it for ovs pod also.



apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  labels:
    app: pdump
  name: pdump
spec:
  selector:
    matchLabels:
      app: pdump
  template:
    metadata:
      labels:
        app: pdump
    spec:
      containers:
        - name: pdump
          image: pdump:latest
          imagePullPolicy: IfNotPresent
          securityContext:
            privileged: true
          volumeMounts:
            - name: var-run-dpdk
              mountPath: /var/run/dpdk
            - name: hugepage
              mountPath: /dev/hugepages-cf
          command: ["sleep", "infinity"]
      volumes:
        - name: var-run-dpdk
          hostPath:
            path: /var/run/dpdk
        - name: hugepage
          hostPath:
            path: /dev/hugepages-cf



Thanks,
Rajesh kumar S R

________________________________
From: Stokes, Ian <ian.stokes at intel.com>
Sent: Wednesday, September 25, 2019 8:19:08 PM
To: Rajesh Kumar; ovs-discuss at openvswitch.org
Cc: anatoly.burakov at intel.com
Subject: Re: [ovs-discuss] OVS - PDUMP: Pdump initialization failure in different container



On 9/16/2019 8:40 AM, Rajesh Kumar wrote:
> Hi,
>=20
> Sorry, Didn't complete my previous mail.
>=20

Hi Rajesh, apologies for the delay on my part in responding, I've been=
=20
out of office the past few weeks.

> The errors I was getting are
> 1)
> root at basepdump-67b4b44448-lt8wf:/# pdump
> EAL: Detected 2 lcore(s)
> EAL: Detected 1 NUMA nodes
> EAL: Multi-process socket /var/run/dpdk/rte/mp_socket_12_19ffcc06b54
> EAL: Probing VFIO support...
> EAL: Cannot initialize tailq: RTE_EVENT_RING
> Tailq 0: qname:<UIO_RESOURCE_LIST>, tqh_first:(nil), tqh_last:0x7fda=
1b17d47c
> Tailq 1: qname:<VFIO_RESOURCE_LIST>, tqh_first:(nil),=20
> tqh_last:0x7fda1b17d4ac
> Tailq 2: qname:<RTE_RING>, tqh_first:0x108064900, tqh_last:0x1080649=
00
> Tailq 3: qname:<RTE_HASH>, tqh_first:(nil), tqh_last:0x7fda1b17d50c
> .............................
> EAL: FATAL: Cannot init tail queues for objects
> EAL: Cannot init tail queues for objects
> PANIC in main():
> Cannot init EAL
> 5: [pdump(+0x2e2a) [0x557832863e2a]]
> 4: [/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xeb)=20
> [0x7fe128dd809b]]
> 3: [pdump(+0x233a) [0x55783286333a]]
> 2: [/usr/lib/x86_64-linux-gnu/librte_eal.so.18.11(__rte_panic+0xbd)=20
> [0x7fe1292b0ca5]]
> 1: [/usr/lib/x86_64-linux-gnu/librte_eal.so.18.11(rte_dump_stack+0x2=
e)=20
> [0x7fe1292c65be]]
> Aborted (core dumped)
>=20
>=20
> 2)
> root at basepdump-67b4b44448-lt8wf:/# pdump
> EAL: Detected 2 lcore(s)
> EAL: Detected 1 NUMA nodes
> PANIC in rte_eal_config_reattach():
> Cannot mmap memory for rte_config at [(nil)], got [0x7ffff...] - ple=
ase=20
> use '--base-virtaddr' option
> 6: [./dpdk-pdump(start+0x2a) [0x5555559c7aa]]
> 5:[/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xeb) [0x7fe128=
dd809b]]
> 4: [./dpdk-pdump(main+0xe2) [0x555555597dd2]]
> 3: [./dpdk-pdump(rte_eal_init+0xc06) [0x555555678416]]
> ..........
> Aborted (core dumped)
>=20

 From the logs above it looks like the secondary process is unable to=20
access the config between the pods. I'm unsure if this is possible=20
myself as I haven't tried this setup before with pdump.

Can I ask if you are specifically sharing the process configs between=20
the pods? Also are you sharing hugepages between the pods and if so,=20
what steps were taken to ensure this?

>=20
> Attached the same errors also.
>=20
> I need in help in figuring out where I'm going wrong.


We'll try to recreate this in or lab setup also as in theory this shou=
ld=20
work.

Regards
Ian

>=20
>=20
>=20
> Thanks,
> Rajesh kumar S R
>=20
>=20
> --------------------------------------------------------------------=
----
> *From:* ovs-discuss-bounces at openvswitch.org=20
> <ovs-discuss-bounces at openvswitch.org> on behalf of Rajesh Kumar=20
> <rajesh.kumar at certesnetworks.com>
> *Sent:* Monday, September 16, 2019 1:00:56 PM
> *To:* ovs-discuss at openvswitch.org
> *Subject:* [ovs-discuss] OVS - PDUMP: Pdump initialization failure i=
n=20
> different container
>=20
> In our kubernetes setup, we are running OVS in a pod with dpdk enabl=
ed.
>=20
> Using 18.11.2.
>=20
> I wanted to use dpdk-pdump as packet capture tool and trying to run=20
> pdump in separate pod.
>=20
> As pdump is a secondary process, it will map to the hugepages alloca=
ted=20
> by primary process (OVS-vswitchd).
>=20
> I'm getting these 2 errors while starting PDUMP as secondary process=
 in=20
> a separate pod.
>=20
>=20
>=20
> Without the container setup, I was able to bringup pdump with OVS
>=20
>=20
>=20
>=20
>=20
> _______________________________________________
> discuss mailing list
> discuss at openvswitch.org
> https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
>=20

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openvswitch.org/pipermail/ovs-discuss/attachments/20190927/09dc6424/attachment.html>


More information about the discuss mailing list