6 systema cimicum delectationis in operatione Kubernetensium [et eorum solutio]

Super annos utendi Kubernetes in productione, multas fabulas iucundas congesserunt quomodo cimices in variis systematis partibus ad injucunda et/vel incomprehensibilia consequentia ad operationem vasorum et siliquae afficientes perduxerunt. In hoc articulo electionem fecimus aliquorum e communium aut iucundorum. Etiamsi numquam felix satis ad eiusmodi casus occurrant, legendo de tam brevibus fabulis detectivis - praesertim "prima-manu" - semper interest, nonne est?..

Fabula 1. Supercronic et Docker tentorium

In una botri periodice accepimus Docker congelatum, quod impediret operationem normalem botri. Eodem tempore haec in Docker acta observata sunt;

level=error msg="containerd: start init process" error="exit status 2: "runtime/cgo: pthread_create failed: No space left on device
SIGABRT: abort
PC=0x7f31b811a428 m=0

goroutine 0 [idle]:

goroutine 1 [running]:
runtime.systemstack_switch() /usr/local/go/src/runtime/asm_amd64.s:252 fp=0xc420026768 sp=0xc420026760
runtime.main() /usr/local/go/src/runtime/proc.go:127 +0x6c fp=0xc4200267c0 sp=0xc420026768
runtime.goexit() /usr/local/go/src/runtime/asm_amd64.s:2086 +0x1 fp=0xc4200267c8 sp=0xc4200267c0

goroutine 17 [syscall, locked to thread]:
runtime.goexit() /usr/local/go/src/runtime/asm_amd64.s:2086 +0x1

…

Quid nos plurimum de hoc errore adiuvat nuntius est: pthread_create failed: No space left on device. Velox Study documentum explicavit Docker non posse processum furca, quam ob causam periodice concrevit.

In vigilantia, sequens imago respondet iis quae aguntur;

Similis condicio in aliis nodis observatur;

Eisdem nodis videmus;

root@kube-node-1 ~ # ps auxfww | grep curl -c
19782
root@kube-node-1 ~ # ps auxfww | grep curl | head
root     16688  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>
root     17398  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>
root     16852  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>
root      9473  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>
root      4664  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>
root     30571  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>
root     24113  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>
root     16475  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>
root      7176  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>
root      1090  0.0  0.0      0     0 ?        Z    Feb06   0:00      |       _ [curl] <defunct>

Contigit quod iste mores consequitur vasculum cum operando supercronic (a Ite utilitas qua utimur ad currendum jobs in siliquis cron);

 _ docker-containerd-shim 833b60bb9ff4c669bb413b898a5fd142a57a21695e5dc42684235df907825567 /var/run/docker/libcontainerd/833b60bb9ff4c669bb413b898a5fd142a57a21695e5dc42684235df907825567 docker-runc
|   _ /usr/local/bin/supercronic -json /crontabs/cron
|       _ /usr/bin/newrelic-daemon --agent --pidfile /var/run/newrelic-daemon.pid --logfile /dev/stderr --port /run/newrelic.sock --tls --define utilization.detect_aws=true --define utilization.detect_azure=true --define utilization.detect_gcp=true --define utilization.detect_pcf=true --define utilization.detect_docker=true
|       |   _ /usr/bin/newrelic-daemon --agent --pidfile /var/run/newrelic-daemon.pid --logfile /dev/stderr --port /run/newrelic.sock --tls --define utilization.detect_aws=true --define utilization.detect_azure=true --define utilization.detect_gcp=true --define utilization.detect_pcf=true --define utilization.detect_docker=true -no-pidfile
|       _ [newrelic-daemon] <defunct>
|       _ [curl] <defunct>
|       _ [curl] <defunct>
|       _ [curl] <defunct>
…

Problema hoc est: cum negotium in supercronico currit, processus ab eo separatur non recte terminareconversus in zombie.

illud: Ut verius dicamus, processus ab operibus magnis spar- tur, sed supercronicus ratio init non est nec processus "capere" potest quos filii sui pariunt. Cum SIGNUM SIGUPUM vel SIGTERM attolluntur, processibus infantis non transmittuntur, unde in processibus puerilibus non terminandis et in zombie statu manentibus. Plura de his omnibus legere potes, e.g talis articulus.

Duo modi sunt ut quaestiones explicet;

1. Pro tempore quo habemus - auge numerum PIDs in systemate uno puncto temporis:

          /proc/sys/kernel/pid_max (since Linux 2.5.34)
                 This file specifies the value at which PIDs wrap around (i.e., the value in this file is one greater than the maximum PID).  PIDs greater than this  value  are  not  allo‐
                 cated;  thus, the value in this file also acts as a system-wide limit on the total number of processes and threads.  The default value for this file, 32768, results in the
                 same range of PIDs as on earlier kernels

2. Vel operas in supercronicas immittendi non directe, sed eodem utendo tinquae potest processus terminare recte et non zombies pariunt.

Fabula 2. "Zombies" cum delendo cgroup

Kubelet multum CPU consumens incepit:

Nemo hoc placet, sic nos armavimus perfectus et quaestionem agere coepit. Eventus investigationis hi sunt:

• Kubelet plusquam tertiam partem suae CPU temporis memoriae datae ab omnibus cgroups trahens;

  

• In nucleo developers' list Mailing invenire potes disputationem de quaestio. Denique ad hoc punctum descendit; variae limae tmpfs et alia similia a systemate non omnino remota Cum delendo cgroup, sic dictum memcg zombie. Serius vel serius e pagina cella delebuntur, sed multum memoria in servo est et nucleus punctum temporis terendo non videt in illis delendis. Quam ob rem ignara servant. Quid hoc etiam accidit? Hic est ministrator cum jobs cron qui constanter novas operas creat, et cum illis siliquas novas. Novae igitur globi creantur pro vasis in eis quae mox deletae sunt.
• Cur cAdvisor in kubelet tantum tempus terit? Hoc facile perspicitur cum simplicissima executione time cat /sys/fs/cgroup/memory/memory.stat. Si in machina sana operatio 0,01 secundis accipit, tunc in cron02 problematico 1,2 secundis accipit. Sermo est cAdvisor, qui notitias e sysfs lentissime legit, rationem memoriae in zombie cgroups considerare conatur.
• Ut zombies fortiter removerent, gazas purgandas conati sunt ut in LKML commendantur: sync; echo 3 > /proc/sys/vm/drop_caches, - sed nucleus magis intricatus evasit et ingruat currus.

Quid facere? Quaestio est constituta (committereet ad descriptionem sedis dimittis nuntius) renovatio nuclei Linux usque ad versionem 4.16.

Historia 3. Systemd et ejus montis

Rursum, kubelet nimis multas facultates in quibusdam nodis consumit, sed hoc tempore nimium memoriam consumit;

Evenit ut problema esset cum systemd adhibito in... Ubuntu 16.04, et hoc fit cum monimenta administrantur quae ad nexum creata sunt. subPath a ConfigMap's or secret's. Post vasculum opus suum complevit the systemd service and its service Mount remain in systemate. Subinde ingens numerus earum cumulatur. Sunt etiam quaestiones de hoc argumento;

1. #5916;
2. kubernetes #57345.

... quorum ultimus refertur ad PR in systemd: #7811 (Exitus in systemd - #7798).

Problema iam non est ibi Ubuntu 18.04, sed si pergere uti vis Ubuntu 16.04, fortasse solutionem nostram de hac re utilem invenies.

Sic DaemonSet hoc fecimus:

---
apiVersion: extensions/v1beta1
kind: DaemonSet
metadata:
  labels:
    app: systemd-slices-cleaner
  name: systemd-slices-cleaner
  namespace: kube-system
spec:
  updateStrategy:
    type: RollingUpdate
  selector:
    matchLabels:
      app: systemd-slices-cleaner
  template:
    metadata:
      labels:
        app: systemd-slices-cleaner
    spec:
      containers:
      - command:
        - /usr/local/bin/supercronic
        - -json
        - /app/crontab
        Image: private-registry.org/systemd-slices-cleaner/systemd-slices-cleaner:v0.1.0
        imagePullPolicy: Always
        name: systemd-slices-cleaner
        resources: {}
        securityContext:
          privileged: true
        volumeMounts:
        - name: systemd
          mountPath: /run/systemd/private
        - name: docker
          mountPath: /run/docker.sock
        - name: systemd-etc
          mountPath: /etc/systemd
        - name: systemd-run
          mountPath: /run/systemd/system/
        - name: lsb-release
          mountPath: /etc/lsb-release-host
      imagePullSecrets:
      - name: antiopa-registry
      priorityClassName: cluster-low
      tolerations:
      - operator: Exists
      volumes:
      - name: systemd
        hostPath:
          path: /run/systemd/private
      - name: docker
        hostPath:
          path: /run/docker.sock
      - name: systemd-etc
        hostPath:
          path: /etc/systemd
      - name: systemd-run
        hostPath:
          path: /run/systemd/system/
      - name: lsb-release
        hostPath:
          path: /etc/lsb-release

... et hoc scriptione utitur;

#!/bin/bash

# we will work only on xenial
hostrelease="/etc/lsb-release-host"
test -f ${hostrelease} && grep xenial ${hostrelease} > /dev/null || exit 0

# sleeping max 30 minutes to dispense load on kube-nodes
sleep $((RANDOM % 1800))

stoppedCount=0
# counting actual subpath units in systemd
countBefore=$(systemctl list-units | grep subpath | grep "run-" | wc -l)
# let's go check each unit
for unit in $(systemctl list-units | grep subpath | grep "run-" | awk '{print $1}'); do
  # finding description file for unit (to find out docker container, who born this unit)
  DropFile=$(systemctl status ${unit} | grep Drop | awk -F': ' '{print $2}')
  # reading uuid for docker container from description file
  DockerContainerId=$(cat ${DropFile}/50-Description.conf | awk '{print $5}' | cut -d/ -f6)
  # checking container status (running or not)
  checkFlag=$(docker ps | grep -c ${DockerContainerId})
  # if container not running, we will stop unit
  if [[ ${checkFlag} -eq 0 ]]; then
    echo "Stopping unit ${unit}"
    # stoping unit in action
    systemctl stop $unit
    # just counter for logs
    ((stoppedCount++))
    # logging current progress
    echo "Stopped ${stoppedCount} systemd units out of ${countBefore}"
  fi
done

... et decurrit singulis 5 minutis utens supradicto supercronico. Dockerfile eius similis est:

FROM ubuntu:16.04
COPY rootfs /
WORKDIR /app
RUN apt-get update && 
    apt-get upgrade -y && 
    apt-get install -y gnupg curl apt-transport-https software-properties-common wget
RUN add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu xenial stable" && 
    curl -fsSL https://download.docker.com/linux/ubuntu/gpg | apt-key add - && 
    apt-get update && 
    apt-get install -y docker-ce=17.03.0*
RUN wget https://github.com/aptible/supercronic/releases/download/v0.1.6/supercronic-linux-amd64 -O 
    /usr/local/bin/supercronic && chmod +x /usr/local/bin/supercronic
ENTRYPOINT ["/bin/bash", "-c", "/usr/local/bin/supercronic -json /app/crontab"]

Fabula 4. Competitiveness cum scheduling siliquae

Animadvertum est: si vasculum in nodo positum habemus et eius imago diutissime exantlaretur, aliud vasculum quod "percutere" idem nodi simpliciter vult non incipit trahere imaginem vasculi novi. Sed expectat donec prioris vasculi imago evulsa sit. Quam ob rem vasculum iam horarium erat et cuius imago receptae poterat in momento minuto modo finietur in statu. containerCreating.

Eventus aliquid simile hoc spectabit;

Normal  Pulling    8m    kubelet, ip-10-241-44-128.ap-northeast-1.compute.internal  pulling image "registry.example.com/infra/openvpn/openvpn:master"

Evenit ut unum imago ex lento subcriptio potest angustos instruere per nodi.

Dolendum quod non multi modi e situ sunt;

1. Conare uti Subcriptio tuo Docker directe in botro vel directe cum botro (exempli gratia, Subcriptio GitLab, Nexus, etc.);
2. Utere utilitates ut Kraken.

Fabula 5. Noctes pendent ob defectum memoriae

Per operationes varias applicationes, etiam condicionem invenimus ubi nodi omnino esse desinunt adiri: SSH non respondet, omnes daemones vigilantes cadunt, et tunc in lignis nihil (vel fere nihil) anomala est.

In picturis dicam tibi utentes exemplo unius nodi ubi MongoDB funguntur.

Hoc est quod supra spectat sicut ad accidentia;

Et sic - после accidentia;

In vigilantia est etiam saltus acutus, in quo nodi desinunt esse in promptu;

Ita ex eenshotsscray manifestum est;

1. Aries in machina prope finem est;
2. Est saltus acutus in RAM phthisi, post quem accessus ad totam machinam abrupte debilis est;
3. Magnum opus in Mongo pervenit, quod processus DBMS magis ad memoriam utendum et actuosius ex orbe legendo cogit.

Evenit ut si in Linux memoria libera deficit (pressio memoriae fit) et nulla permutatio fit, tum ad Cum OOM interfector advenit, actus aequilibrii oriri potest inter paginas proiectas in paginae cella et eas in disco scribens. Hoc fit per kswapd, quod tot paginas memoriae quam maxime ad subsequentem distributionem fortiter liberat.

Infeliciter, magno I/O onere cum parvo liberae memoriae copulato; kswapd fit bottleneck totius systematisquia ligantur ei omnes prouinciis (pagina culparum) in systematis paginarum memoriae. Hoc perquam diu procedere potest si processus memoriam amplius adhibere nolunt, sed in ipso ore OOM interfectoris abyssi figuntur.

Naturalis quaestio est: quare OOM interfector tam sero venit? In current iteratione, OOM interfector perquam stultus est: processus occidet solum cum conatus paginae memoriae collocandae deficit, i.e. si pagina culpa excidit. Hoc non satis diu accidit, quia kswapd paginas memoriae fortiter liberat, dumping paginam cache (totum orbis I/O in systemate, re vera) ad disco. Fusius, cum descriptione graduum ad tales difficultates in nucleo tollendas requiritur, legere potes hic.

Hoc mores ut amplio cum nucleo Linux 4.6.

Fabula 6. Pods adepto adhæsit in pendenti

In nonnullis racemis, in quibus vere multae siliquae operantur, notare incepimus quod plerique "pendent" diutissime in statu. Pendingquamvis Docker continentia ipsa iam in nodis currit et manually elaborari potest.

Praeterea, in describe nihil est mali;

  Type    Reason                  Age                From                     Message
  ----    ------                  ----               ----                     -------
  Normal  Scheduled               1m                 default-scheduler        Successfully assigned sphinx-0 to ss-dev-kub07
  Normal  SuccessfulAttachVolume  1m                 attachdetach-controller  AttachVolume.Attach succeeded for volume "pvc-6aaad34f-ad10-11e8-a44c-52540035a73b"
  Normal  SuccessfulMountVolume   1m                 kubelet, ss-dev-kub07    MountVolume.SetUp succeeded for volume "sphinx-config"
  Normal  SuccessfulMountVolume   1m                 kubelet, ss-dev-kub07    MountVolume.SetUp succeeded for volume "default-token-fzcsf"
  Normal  SuccessfulMountVolume   49s (x2 over 51s)  kubelet, ss-dev-kub07    MountVolume.SetUp succeeded for volume "pvc-6aaad34f-ad10-11e8-a44c-52540035a73b"
  Normal  Pulled                  43s                kubelet, ss-dev-kub07    Container image "registry.example.com/infra/sphinx-exporter/sphinx-indexer:v1" already present on machine
  Normal  Created                 43s                kubelet, ss-dev-kub07    Created container
  Normal  Started                 43s                kubelet, ss-dev-kub07    Started container
  Normal  Pulled                  43s                kubelet, ss-dev-kub07    Container image "registry.example.com/infra/sphinx/sphinx:v1" already present on machine
  Normal  Created                 42s                kubelet, ss-dev-kub07    Created container
  Normal  Started                 42s                kubelet, ss-dev-kub07    Started container

Post aliquot fossiones, posito quod kubelet simpliciter non tempus est mittere omnia informationes de statu siliquarum ac vivacitatis/parati probationes in API servo.

Et postquam auxilium petivimus, parametri sequentes invenimus;

--kube-api-qps - QPS to use while talking with kubernetes apiserver (default 5)
--kube-api-burst  - Burst to use while talking with kubernetes apiserver (default 10) 
--event-qps - If > 0, limit event creations per second to this value. If 0, unlimited. (default 5)
--event-burst - Maximum size of a bursty event records, temporarily allows event records to burst to this number, while still not exceeding event-qps. Only used if --event-qps > 0 (default 10) 
--registry-qps - If > 0, limit registry pull QPS to this value.
--registry-burst - Maximum size of bursty pulls, temporarily allows pulls to burst to this number, while still not exceeding registry-qps. Only used if --registry-qps > 0 (default 10)

Ut videtur, default values ​​​​sunt satis parvaet in 90% tegunt omnes necessitates... Sed in casu nostro hoc non erat satis. Sequentes igitur valores ponemus;

--event-qps=30 --event-burst=40 --kube-api-burst=40 --kube-api-qps=30 --registry-qps=30 --registry-burst=40

... et oculamenta emissa, postquam imaginem sequentem vidimus in graphis vocatuum ad API servo:

... et sic omnia fugere inceperunt!

PS

Ad auxilium suum in colligendis cimicibus et ad hunc articulum componendis, meam gratiam altis exprimo multis fabrum nostrorum comitatus, ac praesertim collegae meo R&D team Andrey Klimentyev (zuzzas).

PPS

Lege etiam in nostro diario:

• «kubectl-debug plugin for debugging in Kubernetes pods".
• Kubernetes apicibus & dolis ansam:
  • «facultates transferentes in botro ad Helm 2 administratione currentes";
  • «De nodi destinatio et onera in applicatione interretiali";
  • «Aditus ad dev situs";
  • «Accelerare bootstrap pro magnis databases".

Source: www.habr.com