Closed. This question is not about programming or software development. It is not currently accepting answers.
This question does not appear to be about a specific programming problem, a software algorithm, or software tools primarily used by programmers. If you believe the question would be on-topic on another Stack Exchange site, you can leave a comment to explain where the question may be able to be answered.
Closed 5 days ago.
Improve this question
Server crashed, when starting the analysis in the logs I found these memory messages.
Would this be relevant to the server's freezer?
[JovaTricolo#xxx log]$ grep -R "Feb 11" messages | egrep -v "audit" | egrep -r "Warning|warning|error|erro|panic|boot|memory"
Feb 11 11:19:56 xxx rsyslogd-2036: error starting up disk queue, using pure in-memory mode [try http://www.rsyslog.com/e/2036 ]
Feb 11 11:19:56 xxx kernel: init_memory_mapping: 0000000000000000-00000000bd2f0000
Feb 11 11:19:56 xxx kernel: init_memory_mapping: 0000000100000000-0000002040000000
Feb 11 11:19:56 xxx kernel: bootmap [0000000000100000 - 0000000000307fff] pages 208
Feb 11 11:19:56 xxx kernel: (8 early reservations) ==> bootmem [0000000000 - 1040000000]
Feb 11 11:19:56 xxx kernel: bootmap [0000001040035000 - 0000001040234fff] pages 200
Feb 11 11:19:56 xxx kernel: (8 early reservations) ==> bootmem [1040000000 - 2040000000]
Feb 11 11:19:56 xxx kernel: Reserving 137MB of memory at 48MB for crashkernel (System RAM: 132096MB)
Feb 11 11:19:56 xxx kernel: PM: Registered nosave memory: 000000000009c000 - 0000000000100000
Feb 11 11:19:56 xxx kernel: PM: Registered nosave memory: 00000000bd2f0000 - 00000000bd31c000
Feb 11 11:19:56 xxx kernel: PM: Registered nosave memory: 00000000bd31c000 - 00000000bd35b000
Feb 11 11:19:56 xxx kernel: PM: Registered nosave memory: 00000000bd35b000 - 00000000c0000000
Feb 11 11:19:56 xxx kernel: PM: Registered nosave memory: 00000000c0000000 - 00000000e0000000
Feb 11 11:19:56 xxx kernel: PM: Registered nosave memory: 00000000e0000000 - 00000000f0000000
Feb 11 11:19:56 xxx kernel: PM: Registered nosave memory: 00000000f0000000 - 00000000fe000000
Feb 11 11:19:56 xxx kernel: PM: Registered nosave memory: 00000000fe000000 - 0000000100000000
Feb 11 11:19:56 xxx kernel: please try 'cgroup_disable=memory' option if you don't want memory cgroups
Feb 11 11:19:56 xxx kernel: Initializing cgroup subsys memory
Feb 11 11:19:56 xxx kernel: Freeing initrd memory: 15695k freed
Feb 11 11:19:56 xxx kernel: Non-volatile memory driver v1.3
Feb 11 11:19:56 xxx kernel: crash memory driver: version 1.1
Feb 11 11:19:56 xxx kernel: Freeing unused kernel memory: 1252k freed
Feb 11 11:19:56 xxx kernel: Freeing unused kernel memory: 1051k freed
Feb 11 11:19:56 xxx kernel: Freeing unused kernel memory: 1734k freed
Feb 11 11:19:56 xxx kernel: EXT4-fs (dm-13): warning: checktime reached, running e2fsck is recommended
Feb 11 11:19:56 xxx kernel: EXT4-fs (dm-14): warning: checktime reached, running e2fsck is recommended
Feb 11 11:19:56 xxx kernel: EXT4-fs (dm-15): warning: checktime reached, running e2fsck is recommended
Feb 11 11:19:56 xxx kernel: EXT4-fs (dm-16): warning: maximal mount count reached, running e2fsck is recommended
Feb 11 11:50:05 xxx snmpd[24795]: Warning: no access control information configured.#012 It's unlikely this agent can serve any useful purpose in this state.#012 Run "snmpconf -g basic_setup" to help you configure the snmpd.conf file for this agent.
[JovaTricolo#xxx log]$
Is there any solution for this?
We have kubernetes cluster, and we are running jenkins in it. Our jenkins restart after every 48 hours, when we check the kubelet logs for that worker where jenkins deployed, it gives error
Feb 15 14:52:01 myworker kernel: Memory cgroup out of memory: Kill process 110129 (Computer.thread) score 1972 or sacrifice child
Feb 15 14:52:01 myworker kernel: Killed process 50179 (java), UID 1000, total-vm:17378260kB, anon-rss:8371056kB, file-rss:29676kB, shmem-rss:0kB
where 50179 is java process for jenkins.
We set limit in kubernetes for jenkins as 8Gi
resources:
limits:
cpu: 3500m
memory: 8Gi
requests:
cpu: "1"
memory: 4Gi
I also check newrelic alerts, which we integrated with our pods, it never goes beyond 5GB in memory.
Details logs below.
Feb 15 14:52:01 myworker kernel: Download metada cpuset=kubepods-burstable-pod1840326e_dca6_4e8c_a55a_f4fb9a7c95fa.slice:cri-containerd:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d mems_allowed=0
Feb 15 14:52:01 myworker kernel: CPU: 6 PID: 115222 Comm: Download metada Kdump: loaded Tainted: G ------------ T 3.10.0-1160.15.2.el7.x86_64 #1
Feb 15 14:52:01 myworker kernel: Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 12/12/2018
Feb 15 14:52:01 myworker kernel: Call Trace:
Feb 15 14:52:01 myworker kernel: [<ffffffff82581fba>] dump_stack+0x19/0x1b
Feb 15 14:52:01 myworker kernel: [<ffffffff8257c8da>] dump_header+0x90/0x229
Feb 15 14:52:01 myworker kernel: [<ffffffff8209d378>] ? ep_poll_callback+0xf8/0x220
Feb 15 14:52:01 myworker kernel: [<ffffffff81fc1d16>] ? find_lock_task_mm+0x56/0xc0
Feb 15 14:52:01 myworker kernel: [<ffffffff8203caa8>] ? try_get_mem_cgroup_from_mm+0x28/0x60
Feb 15 14:52:01 myworker kernel: [<ffffffff81fc227d>] oom_kill_process+0x2cd/0x490
Feb 15 14:52:01 myworker kernel: [<ffffffff82040ebc>] mem_cgroup_oom_synchronize+0x55c/0x590
Feb 15 14:52:01 myworker kernel: [<ffffffff82040320>] ? mem_cgroup_charge_common+0xc0/0xc0
Feb 15 14:52:01 myworker kernel: [<ffffffff81fc2b64>] pagefault_out_of_memory+0x14/0x90
Feb 15 14:52:01 myworker kernel: [<ffffffff8257ade6>] mm_fault_error+0x6a/0x157
Feb 15 14:52:01 myworker kernel: [<ffffffff8258f8d1>] __do_page_fault+0x491/0x500
Feb 15 14:52:01 myworker kernel: [<ffffffff8258f975>] do_page_fault+0x35/0x90
Feb 15 14:52:01 myworker kernel: [<ffffffff8258b778>] page_fault+0x28/0x30
Feb 15 14:52:01 myworker kernel: Task in /system.slice/containerd.service/kubepods-burstable-pod1840326e_dca6_4e8c_a55a_f4fb9a7c95fa.slice:cri-containerd:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d killed as a result of limit of /system.slice/containerd.service/kubepods-burstable-pod1840326e_dca6_4e8c_a55a_f4fb9a7c95fa.slice:cri-containerd:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d
Feb 15 14:52:01 myworker kernel: memory: usage 8388608kB, limit 8388608kB, failcnt 111634
Feb 15 14:52:01 myworker kernel: memory+swap: usage 8388608kB, limit 9007199254740988kB, failcnt 0
Feb 15 14:52:01 myworker kernel: kmem: usage 0kB, limit 9007199254740988kB, failcnt 0
Feb 15 14:52:01 myworker kernel: Memory cgroup stats for /system.slice/containerd.service/kubepods-burstable-pod1840326e_dca6_4e8c_a55a_f4fb9a7c95fa.slice:cri-containerd:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d: cache:20KB rss:8388588KB rss_huge:6144KB mapped_file:0KB swap:0KB inactive_anon:0KB active_anon:8388556KB inactive_file:4KB active_file:0KB unevictable:0KB
Feb 15 14:52:01 myworker kernel: [ pid ] uid tgid total_vm rss nr_ptes swapents oom_score_adj name
[root#myworker log]# head messages -n376428 | tail -n 40
Feb 15 14:52:01 myworker kernel: [115493] 1000 115493 2059 462 8 0 969 git
Feb 15 14:52:01 myworker kernel: [115497] 1000 115497 1764 350 8 0 969 git
Feb 15 14:52:01 myworker kernel: [115498] 1000 115498 24351 2784 17 0 969 git-remote-http
Feb 15 14:52:01 myworker kernel: Memory cgroup out of memory: Kill process 115496 (git fetch --tag) score 1972 or sacrifice child
Feb 15 14:52:01 myworker kernel: Killed process 115493 (git), UID 1000, total-vm:8236kB, anon-rss:296kB, file-rss:1552kB, shmem-rss:0kB
Feb 15 14:52:01 myworker containerd: time="2022-02-15T14:52:01.791126760Z" level=info msg="TaskOOM event &TaskOOM{ContainerID:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d,XXX_unrecognized:[],}"
Feb 15 14:52:01 myworker kernel: Download metada invoked oom-killer: gfp_mask=0xd0, order=0, oom_score_adj=969
Feb 15 14:52:01 myworker kernel: Download metada cpuset=kubepods-burstable-pod1840326e_dca6_4e8c_a55a_f4fb9a7c95fa.slice:cri-containerd:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d mems_allowed=0
Feb 15 14:52:01 myworker kernel: CPU: 6 PID: 115222 Comm: Download metada Kdump: loaded Tainted: G ------------ T 3.10.0-1160.15.2.el7.x86_64 #1
Feb 15 14:52:01 myworker kernel: Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 12/12/2018
Feb 15 14:52:01 myworker kernel: Call Trace:
Feb 15 14:52:01 myworker kernel: [<ffffffff82581fba>] dump_stack+0x19/0x1b
Feb 15 14:52:01 myworker kernel: [<ffffffff8257c8da>] dump_header+0x90/0x229
Feb 15 14:52:01 myworker kernel: [<ffffffff8209d378>] ? ep_poll_callback+0xf8/0x220
Feb 15 14:52:01 myworker kernel: [<ffffffff81fc1d16>] ? find_lock_task_mm+0x56/0xc0
Feb 15 14:52:01 myworker kernel: [<ffffffff8203caa8>] ? try_get_mem_cgroup_from_mm+0x28/0x60
Feb 15 14:52:01 myworker kernel: [<ffffffff81fc227d>] oom_kill_process+0x2cd/0x490
Feb 15 14:52:01 myworker kernel: [<ffffffff82040ebc>] mem_cgroup_oom_synchronize+0x55c/0x590
Feb 15 14:52:01 myworker kernel: [<ffffffff82040320>] ? mem_cgroup_charge_common+0xc0/0xc0
Feb 15 14:52:01 myworker kernel: [<ffffffff81fc2b64>] pagefault_out_of_memory+0x14/0x90
Feb 15 14:52:01 myworker kernel: [<ffffffff8257ade6>] mm_fault_error+0x6a/0x157
Feb 15 14:52:01 myworker kernel: [<ffffffff8258f8d1>] __do_page_fault+0x491/0x500
Feb 15 14:52:01 myworker kernel: [<ffffffff8258f975>] do_page_fault+0x35/0x90
Feb 15 14:52:01 myworker kernel: [<ffffffff8258b778>] page_fault+0x28/0x30
Feb 15 14:52:01 myworker kernel: Task in /system.slice/containerd.service/kubepods-burstable-pod1840326e_dca6_4e8c_a55a_f4fb9a7c95fa.slice:cri-containerd:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d killed as a result of limit of /system.slice/containerd.service/kubepods-burstable-pod1840326e_dca6_4e8c_a55a_f4fb9a7c95fa.slice:cri-containerd:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d
Feb 15 14:52:01 myworker kernel: memory: usage 8388608kB, limit 8388608kB, failcnt 111634
Feb 15 14:52:01 myworker kernel: memory+swap: usage 8388608kB, limit 9007199254740988kB, failcnt 0
Feb 15 14:52:01 myworker kernel: kmem: usage 0kB, limit 9007199254740988kB, failcnt 0
Feb 15 14:52:01 myworker kernel: Memory cgroup stats for /system.slice/containerd.service/kubepods-burstable-pod1840326e_dca6_4e8c_a55a_f4fb9a7c95fa.slice:cri-containerd:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d: cache:20KB rss:8388588KB rss_huge:6144KB mapped_file:0KB swap:0KB inactive_anon:0KB active_anon:8388556KB inactive_file:4KB active_file:0KB unevictable:0KB
Feb 15 14:52:01 myworker kernel: [ pid ] uid tgid total_vm rss nr_ptes swapents oom_score_adj name
Feb 15 14:52:01 myworker kernel: [41710] 1000 41710 285 1 4 0 969 tini
Feb 15 14:52:01 myworker kernel: [50179] 1000 50179 4344565 2100159 4662 0 969 java
Feb 15 14:52:01 myworker kernel: [115497] 1000 115497 1764 350 8 0 969 git
Feb 15 14:52:01 myworker kernel: [115498] 1000 115498 24351 2784 17 0 969 git-remote-http
Feb 15 14:52:01 myworker kernel: Memory cgroup out of memory: Kill process 110129 (Computer.thread) score 1972 or sacrifice child
Feb 15 14:52:01 myworker kernel: Killed process 50179 (java), UID 1000, total-vm:17378260kB, anon-rss:8371056kB, file-rss:29676kB, shmem-rss:0kB
Feb 15 14:52:03 myworker containerd: time="2022-02-15T14:52:03.132654815Z" level=info msg="Finish piping stderr of container \"7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d\""
Feb 15 14:52:03 myworker containerd: time="2022-02-15T14:52:03.132676088Z" level=info msg="Finish piping stdout of container \"7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d\""
Feb 15 14:52:03 myworker containerd: time="2022-02-15T14:52:03.134738144Z" level=info msg="TaskExit event &TaskExit{ContainerID:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d,ID:7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d,Pid:41710,ExitStatus:137,ExitedAt:2022-02-15 14:52:03.134458495 +0000 UTC,XXX_unrecognized:[],}"
Feb 15 14:52:03 myworker containerd: time="2022-02-15T14:52:03.248040140Z" level=info msg="shim disconnected" id=7fc11e70ccd4fd078b8d243f2710ecc1404955bf52a5cb05eb54f2917086420d
Only problem I can see here is, we are telling kubernetes to go till 8Gb but Memory cgroup might have limit below 8Gb and when it try to reach something beyond 5Gb it kill the pod and it restart again.
What is the best way to know the Memory cgroup limit? and is there way to know which pods/process are using this cgroup?
Questions:
Q1: What kind of cluster do you use? Minikube, kubeadm or managed by cloud GKE, EKS, AKS?
A1: kubeadm
Q2: Which version of kubernetes do you use?
A2: v1.21.3
Q3: From when the problem with restart jenkins pod has been started?
A3: Issue might be from the beginning, but we start noticing recently when we moved more jobs to kubernetes based jenkins.
Q4: Could you paste an output from jenkins pods using kubectl describe pod ?
A4:
# kubectl describe pod -n jenkins jenkins-jenkins-instance
Name: jenkins-jenkins-instance
Namespace: jenkins
Priority: 0
Node: myworker/192.168.X.X
Start Time: Sun, 13 Mar 2022 15:12:19 +0000
Labels: app=jenkins-operator
jenkins-cr=jenkins-instance
Annotations: <none>
Status: Running
IP: 192.168.113.152
IPs:
IP: 192.168.113.152
Controlled By: Jenkins/jenkins-instance
Containers:
jenkins-master:
Container ID: containerd://70e68b7b069404f825b53e9d8f0dac22c595074e5bdc4659cae5248e25af8e00
Image: jenkins/jenkins:lts
Image ID: docker.io/jenkins/jenkins#sha256:b414f82151b865d3efd49ec27a944f624188d09fec58700cddfbe6bae2450f77
Ports: 8080/TCP, 50000/TCP
Host Ports: 0/TCP, 0/TCP
Command:
bash
-c
/var/jenkins/scripts/init.sh && exec /sbin/tini -s -- /usr/local/bin/jenkins.sh
State: Running
Started: Sun, 13 Mar 2022 15:12:20 +0000
Ready: True
Restart Count: 0
Limits:
cpu: 3500m
memory: 8Gi
Requests:
cpu: 1
memory: 4Gi
Liveness: http-get http://:http/login delay=100s timeout=5s period=10s #success=1 #failure=12
Readiness: http-get http://:http/login delay=80s timeout=1s period=10s #success=1 #failure=10
Environment:
COPY_REFERENCE_FILE_LOG: /var/lib/jenkins/copy_reference_file.log
NEW_RELIC_METADATA_KUBERNETES_CLUSTER_NAME: IAD.Prod
NEW_RELIC_METADATA_KUBERNETES_NODE_NAME: (v1:spec.nodeName)
NEW_RELIC_METADATA_KUBERNETES_NAMESPACE_NAME: jenkins (v1:metadata.namespace)
NEW_RELIC_METADATA_KUBERNETES_POD_NAME: jenkins-jenkins-instance (v1:metadata.name)
NEW_RELIC_METADATA_KUBERNETES_CONTAINER_NAME: master
NEW_RELIC_METADATA_KUBERNETES_CONTAINER_IMAGE_NAME: jenkins/jenkins:lts
JAVA_OPTS: -XX:MinRAMPercentage=50.0 -XX:MaxRAMPercentage=80.0 -Djenkins.install.runSetupWizard=false -Djava.awt.headless=true
JENKINS_HOME: /var/lib/jenkins
Mounts:
/var/jenkins/init-configuration from init-configuration (ro)
/var/jenkins/operator-credentials from operator-credentials (ro)
/var/jenkins/scripts from scripts (ro)
/var/lib/jenkins from jenkins-home (rw)
/var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-fc57k (ro)
Conditions:
Type Status
Initialized True
Ready True
ContainersReady True
PodScheduled True
Volumes:
jenkins-home:
Type: EmptyDir (a temporary directory that shares a pod's lifetime)
Medium:
SizeLimit: <unset>
scripts:
Type: ConfigMap (a volume populated by a ConfigMap)
Name: jenkins-operator-scripts-jenkins-instance
Optional: false
init-configuration:
Type: ConfigMap (a volume populated by a ConfigMap)
Name: jenkins-operator-init-configuration-jenkins-instance
Optional: false
operator-credentials:
Type: Secret (a volume populated by a Secret)
SecretName: jenkins-operator-credentials-jenkins-instance
Optional: false
kube-api-access-fc57k:
Type: Projected (a volume that contains injected data from multiple sources)
TokenExpirationSeconds: 3607
ConfigMapName: kube-root-ca.crt
ConfigMapOptional: <nil>
DownwardAPI: true
QoS Class: Burstable
Node-Selectors: <none>
Tolerations: node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events: <none>
Q5: How we deploy jenkins?
A5: We are using Jenkins-operator to deploy jenkins.
I have a cluster of 3 kafka brokers Version 0.10.2.1. Each broker has it's own host 2 cpu / 16G RAM, In addition we are using docker to wrap the broker process.
The problems is as follows:
Almost every day at the same time we see all of our kafka clients failed for 10 minutes.
At the beginning I thought it is related to Kafka No broker in ISR for partition
But after a while I discover that the broker just crash due to OOM-killer.
I also played with the Xmx and Xms before I discover that it is the OOM-killer. I had:
-Xmx2048M -Xms2048M
-Xmx4096M -Xms2048M
Same behavior for both
In addition currently we don't have ulimit
>> ulimit
unlimited
less kern.log
LOGS:
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761019] run-parts invoked oom-killer: gfp_mask=0x26000c0, order=2, oom_score_adj=0
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761022] run-parts cpuset=/ mems_allowed=0
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761026] CPU: 1 PID: 12266 Comm: run-parts Not tainted 4.4.0-59-generic #80-Ubuntu
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761027] Hardware name: Xen HVM domU, BIOS 4.2.amazon 02/16/2017
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761029] 0000000000000286 000000004811d7da ffff880036967af0 ffffffff813f7583
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761031] ffff880036967cc8 ffff880439f2f000 ffff880036967b60 ffffffff8120ad5e
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761033] ffffffff81cd2dc7 0000000000000000 ffffffff81e67760 0000000000000206
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761036] Call Trace:
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761041] [<ffffffff813f7583>] dump_stack+0x63/0x90
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761044] [<ffffffff8120ad5e>] dump_header+0x5a/0x1c5
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761048] [<ffffffff81192722>] oom_kill_process+0x202/0x3c0
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761049] [<ffffffff81192b49>] out_of_memory+0x219/0x460
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761052] [<ffffffff81198abd>] __alloc_pages_slowpath.constprop.88+0x8fd/0xa70
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761054] [<ffffffff81198eb6>] __alloc_pages_nodemask+0x286/0x2a0
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761056] [<ffffffff81198f6b>] alloc_kmem_pages_node+0x4b/0xc0
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761060] [<ffffffff8107ea5e>] copy_process+0x1be/0x1b70
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761063] [<ffffffff81391bcc>] ? apparmor_file_alloc_security+0x5c/0x220
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761066] [<ffffffff811ed05a>] ? kmem_cache_alloc+0x1ca/0x1f0
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761070] [<ffffffff81347bd3>] ? security_file_alloc+0x33/0x50
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761073] [<ffffffff810caf11>] ? __raw_callee_save___pv_queued_spin_unlock+0x11/0x20
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761074] [<ffffffff810805a0>] _do_fork+0x80/0x360
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761076] [<ffffffff81080929>] SyS_clone+0x19/0x20
Jan 23 06:25:16 kafka10-172-40-103-177 kernel: [16504862.761080] [<ffffffff818384f2>] entry_SYSCALL_64_fastpath+0x16/0x71
And ....
Jan 24 06:25:25 kafka10-172-40-103-177 kernel: [16591270.954463] Out of memory: Kill process 16123 (java) score 134 or sacrifice child
Jan 24 06:25:25 kafka10-172-40-103-177 kernel: [16591270.958609] Killed process 16123 (java) total-vm:11977548kB, anon-rss:2035780kB, file-rss:67848kB
Any suggestion of how to approach this ??
We found the problem.
First I will say that adding more RAM to the machine also solved the problem but it is "expensive solution".
The problem was as follows:
Since I was working with EC2 ubuntu distribution I got daily crontabs in all of my cluster exactly at the same time. One of the scripts was mlocate this script apparently took too many resources.
I assume that since all cluster of kafka has some issues with IO and Memory, brokers was trying to use more memory and then the OOM killer killed them.
When 2 of my 3 brokers were down some services were down.
So the solution was:
Change the crontab to work in different hours of the day in each
broker.
Disable mlocate
I also faced the same issue below mentioned blog helped me out :
https://docs.confluent.io/current/kafka/deployment.html
How to decide Kafka Cluster size
https://community.hortonworks.com/articles/80813/kafka-best-practices-1.html
And please make sure that the swap is enabled on all the brokers.
We can issue a host bus scan on Linux host to get the /dev/sd* devices on host.
The scan is issued using this command :
echo "- - -" > /sys/class/scsi_host/host0/scan
Can someone please tell me the flow of events in the Linux userspace/kernel code which leads to formation of /dev/sd* devices post scsi scan ?
Is this a PCI bus scan OR SCSI commands sent to controller of storage OR something new ?
Looked upon the code how iscsid scan works. __scsi_scan_target is the key function which does the SCSI scan.
This function first it tries to probe and add LUN 0 using scsi_probe_and_add_lun function.
Dec 24 03:24:28 localhost kernel: [<ffffffff813e8c94>] scsi_probe_and_add_lun+0x3f4/0xc80
Dec 24 03:24:28 localhost kernel: [<ffffffff815e5a51>] ? printk+0x77/0x8e
Dec 24 03:24:28 localhost kernel: [<ffffffff813e9a0e>] __scsi_scan_target+0x13e/0x270
Dec 24 03:24:28 localhost kernel: [<ffffffff813e9c30>] scsi_scan_target+0xf0/0x110
Dec 24 03:24:28 localhost kernel: [<ffffffffa050eedf>] iscsi_user_scan_session.part.13+0x10f/0x150 [scsi_transport_iscsi]
Dec 24 03:24:28 localhost kernel: [<ffffffffa050ef20>] ? iscsi_user_scan_session.part.13+0x150/0x150 [scsi_transport_iscsi]
Dec 24 03:24:28 localhost kernel: [<ffffffffa050ef41>] iscsi_user_scan_session+0x21/0x30 [scsi_transport_iscsi]
Dec 24 03:24:28 localhost kernel: [<ffffffff813b1b63>] device_for_each_child+0x53/0x90
Dec 24 03:24:28 localhost kernel: [<ffffffffa050cb6c>] iscsi_user_scan+0x3c/0x60 [scsi_transport_iscsi]
Dec 24 03:24:28 localhost kernel: [<ffffffff813eb835>] store_scan+0xa5/0x100
Dec 24 03:24:28 localhost kernel: [<ffffffff813b1138>] dev_attr_store+0x18/0x30
Dec 24 03:24:28 localhost kernel: [<ffffffff81225686>] sysfs_write_file+0xc6/0x140
Dec 24 03:24:28 localhost kernel: [<ffffffff811afa7d>] vfs_write+0xbd/0x1e0
Dec 24 03:24:28 localhost kernel: [<ffffffff811b04c8>] SyS_write+0x58/0xb0
Dec 24 03:24:28 localhost kernel: [<ffffffff815fc9d9>] system_call_fastpath+0x16/0x1b
scsi_probe_and_add_lun function will :
1. allocate scsi_device data structure
2. call scsi_probe_lun and send SCSI enquiries to LUN 0.
3. call scsi_add_lun to fill the scsi_device data structure from the enquiries done.
this data structure is added using scsi_sysfs_add_sdev function.
Then it tries to send SCSI report LUN command to get the number of LUNS in function scsi_report_lun_scan. This function finds the number of LUNs and again tries to add all the LUNs using scsi_probe_and_add_lun function.
scsi_sysfs_add_sdev function adds the devices by calling scsi_target_add and device_add for scsi_device.
This device_add calls the drivers' probe function (sd_probe):
Dec 24 03:24:28 localhost kernel: [<ffffffffa01d77d0>] sd_probe+0x320/0x380 [sd_mod]
Dec 24 03:24:28 localhost kernel: [<ffffffff813b6917>] driver_probe_device+0x87/0x390
Dec 24 03:24:28 localhost kernel: [<ffffffff813b6c20>] ? driver_probe_device+0x390/0x390
Dec 24 03:24:28 localhost kernel: [<ffffffff813b6c5b>] __device_attach+0x3b/0x40
Dec 24 03:24:28 localhost kernel: [<ffffffff813b477b>] bus_for_each_drv+0x6b/0xb0
Dec 24 03:24:28 localhost kernel: [<ffffffff813b6818>] device_attach+0x88/0xa0
Dec 24 03:24:28 localhost kernel: [<ffffffff813b5b18>] bus_probe_device+0x98/0xc0
Dec 24 03:24:28 localhost kernel: [<ffffffff813b3584>] device_add+0x4c4/0x7a0
Dec 24 03:24:28 localhost kernel: [<ffffffff813c2ebc>] ? __pm_runtime_resume+0x5c/0x80
Dec 24 03:24:28 localhost kernel: [<ffffffff813eba9c>] scsi_sysfs_add_sdev+0xac/0x320
Dec 24 03:24:28 localhost kernel: [<ffffffff813e9317>] scsi_probe_and_add_lun+0xa77/0xc80
Dec 24 03:24:28 localhost kernel: [<ffffffff815eceff>] scsi_report_lun_scan+0x39a/0x5f1
Dec 24 03:24:28 localhost kernel: [<ffffffff813e9a24>] __scsi_scan_target+0x154/0x270
Dec 24 03:24:28 localhost kernel: [<ffffffff813e9c30>] scsi_scan_target+0xf0/0x110
Dec 24 03:24:28 localhost kernel: [<ffffffffa050eedf>] iscsi_user_scan_session.part.13+0x10f/0x150 [scsi_transport_iscsi]
Dec 24 03:24:28 localhost kernel: [<ffffffffa050ef20>] ? iscsi_user_scan_session.part.13+0x150/0x150 [scsi_transport_iscsi]
Dec 24 03:24:28 localhost kernel: [<ffffffffa050ef41>] iscsi_user_scan_session+0x21/0x30 [scsi_transport_iscsi]
Dec 24 03:24:28 localhost kernel: [<ffffffff813b1b63>] device_for_each_child+0x53/0x90
Dec 24 03:24:28 localhost kernel: [<ffffffffa050cb6c>] iscsi_user_scan+0x3c/0x60 [scsi_transport_iscsi]
Dec 24 03:24:28 localhost kernel: [<ffffffff813eb835>] store_scan+0xa5/0x100
Dec 24 03:24:28 localhost kernel: [<ffffffff813b1138>] dev_attr_store+0x18/0x30
Dec 24 03:24:28 localhost kernel: [<ffffffff81225686>] sysfs_write_file+0xc6/0x140
Dec 24 03:24:28 localhost kernel: [<ffffffff811afa7d>] vfs_write+0xbd/0x1e0
Dec 24 03:24:28 localhost kernel: [<ffffffff811b04c8>] SyS_write+0x58/0xb0
Dec 24 03:24:28 localhost kernel: [<ffffffff815fc9d9>] system_call_fastpath+0x16/0x1b
linux-3.14.69\drivers\scsi\sd.c
This sd_probe function allocates the disk and launches thread sd_probe_async.
async_schedule_domain(sd_probe_async, sdkp, &scsi_sd_probe_domain);
sd_probe_async calls add_disk and prints the message :
"Attached SCSI %sdisk"
I'm trying to understand the output from the backtrace I captured using kprobes. If you'd like to see the full stack trace, that's fine, but it's not necessary for the question. Below is an excerpt:
Jul 24 16:31:34 scilinx kernel: [<ffffffff813a2b2e>] ? ata_scsi_port_error_handler+0x4be/0x710
Jul 24 16:31:34 scilinx kernel: [<ffffffff813a2ea8>] ? ata_scsi_cmd_error_handler+0x128/0x180
Jul 24 16:31:34 scilinx kernel: [<ffffffff813a2f98>] ? ata_scsi_error+0x98/0xd0
Jul 24 16:31:34 scilinx kernel: [<ffffffff81386cfa>] ? scsi_error_handler+0x12a/0x810
Jul 24 16:31:34 scilinx kernel: [<ffffffff81386bd0>] ? scsi_error_handler+0x0/0x810
Jul 24 16:31:34 scilinx kernel: [<ffffffff8109aef6>] ? kthread+0x96/0xa0
You'll notice that scsi_error_handler is adjacent to itself in the call stack but I cannot find why this is. Here is the scsi_error_handler function for this kernel. As you can see, it does not call itself. So why does the stacktrace show it adjacent to itself like this?
Thanks.