![6 ensèk sistèm amizan nan operasyon Kubernetes [ak solisyon yo]](https://prohoster.info/wp-content/plugins/wp-fastest-cache-premium/pro/images/blank.gif)
Pandan ane yo nan sèvi ak Kubernetes nan pwodiksyon, nou te akimile anpil istwa enteresan sou ki jan pinèz nan divès eleman sistèm mennen nan konsekans dezagreyab ak / oswa enkonpreyansib ki afekte operasyon an nan resipyan ak gous. Nan atik sa a nou te fè yon seleksyon nan kèk nan sa yo ki pi komen oswa enteresan. Menm si ou pa janm gen chans pou rankontre sitiyasyon sa yo, li sou istwa kout detektif sa yo - espesyalman "premye men" - toujou enteresan, pa vre?...
Istwa 1. Supercronic ak Docker pandye
Sou youn nan grap yo, nou te resevwa detanzantan yon Docker nan frizè, ki te entèfere ak fonksyone nòmal gwoup la. An menm tan an, sa ki annapre yo te obsève nan mòso Docker yo:
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
…
Sa ki pi enterese nou nan erè sa a se mesaj la: pthread_create failed: No space left on device. Etid rapid eksplike ke Docker pa t 'kapab fouchèt yon pwosesis, ki se poukisa li detanzantan jele.
Nan siveyans, foto sa a koresponn ak sa k ap pase:
Yon sitiyasyon ki sanble obsève sou lòt nœuds:
Nan nœuds yo menm nou wè:
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>Li te tounen soti ke konpòtman sa a se yon konsekans gous la ap travay ak (yon sèvis piblik Go ke nou itilize pou kouri travay cron nan gous):
_ 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>
…Pwoblèm nan se sa a: lè yon travay kouri nan supercronic, pwosesis la anjandre pa li pa ka sispann kòrèkteman, vire nan .
Note: Pou yo ka pi presi, pwosesis yo anjandre pa travay cron, men supercronic se pa yon sistèm init epi li pa ka "adopte" pwosesis ke pitit li yo anjandre. Lè siyal SIGHUP oswa SIGTERM yo leve, yo pa transmèt yo nan pwosesis timoun yo, sa ki lakòz pwosesis timoun yo pa sispann epi rete nan estati zonbi. Ou ka li plis sou tout bagay sa yo, pou egzanp, nan .
Gen yon koup nan fason yo rezoud pwoblèm:
- Kòm yon solisyon tanporè - ogmante kantite PID nan sistèm nan nan yon sèl pwen nan tan:
/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 - Oswa lanse travay nan supercronic pa dirèkteman, men lè l sèvi avèk menm bagay la , ki se kapab mete fen nan pwosesis kòrèkteman epi yo pa anjandre zonbi.
Istwa 2. "Zonbi" lè w ap efase yon gwoup c
Kubelet te kòmanse konsome anpil CPU:
Pa gen moun ki pral renmen sa, se konsa nou ame tèt nou epi yo te kòmanse fè fas ak pwoblèm nan. Rezilta yo nan ankèt la te jan sa a:
- Kubelet depanse plis pase yon tyè nan tan CPU li yo rale done memwa nan tout cgroups:
- Nan lis adrès devlopè nwayo yo ou ka jwenn . Nan ti bout tan, pwen an desann nan sa a: divès dosye tmpfs ak lòt bagay ki sanble yo pa konplètman retire nan sistèm nan lè efase yon cgroup, sa yo rele zonbi. Talè oswa pita yo pral efase nan kachèt paj la, men gen anpil memwa sou sèvè a ak nwayo a pa wè pwen nan gaspiye tan sou efase yo. Se poutèt sa yo kontinye anpile. Poukisa sa a menm rive? Sa a se yon sèvè ki gen djòb cron ki toujou ap kreye nouvo djòb, epi avèk yo nouvo gous. Kidonk, nouvo cgroups yo kreye pou resipyan ki ladan yo, ki byento efase.
- Poukisa cAdvisor nan kubelet gaspiye anpil tan? Sa a se fasil wè ak ekzekisyon ki pi senp
time cat /sys/fs/cgroup/memory/memory.stat. Si sou yon machin ki an sante operasyon an pran 0,01 segonn, Lè sa a, sou pwoblèm nan cron02 li pran 1,2 segonn. Bagay la se ke cAdvisor, ki li done ki soti nan sysfs trè dousman, eseye pran an kont memwa yo itilize nan cgroup zonbi. - Pou retire zonbi ak fòs, nou te eseye netwaye kachèt jan yo rekòmande nan LKML:
sync; echo 3 > /proc/sys/vm/drop_caches, - men nwayo a te vin pi konplike ak fè aksidan machin nan.
Kisa pou fe? Pwoblèm nan ap rezoud (, ak pou yon deskripsyon gade ) mete ajou nwayo Linux la nan vèsyon 4.16.
Istwa 3. Systemd ak mòn li yo
Ankò, kubelet la ap konsome twòp resous sou kèk nœuds, men fwa sa a li konsome twòp memwa:
Li te tounen soti ke gen yon pwoblèm nan systemd yo itilize nan Ubuntu 16.04, epi li rive lè jere mòn yo kreye pou koneksyon. subPath soti nan ConfigMap oswa sekrè a. Apre gous la fini travay li sèvis systemd la ak mòn sèvis li yo rete nan sistèm. Apre yon tan, yon gwo kantite nan yo akimile. Gen menm pwoblèm sou sijè sa a:
- ;
- .
... dènye a nan ki refere a PR a nan systemd: (pwoblèm nan systemd - ).
Pwoblèm nan pa egziste ankò nan Ubuntu 18.04, men si ou vle kontinye itilize Ubuntu 16.04, ou ka jwenn solisyon nou an itil sou sijè sa a.
Se konsa, nou te fè DaemonSet sa a:
---
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... epi li itilize script sa a:
#!/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... epi li kouri chak 5 minit lè l sèvi avèk supercronic mansyone deja. Dockerfile li yo sanble sa a:
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"]Istwa 4. Konpetitivite lè w ap planifye gous yo
Li te remake ke: si nou gen yon gous mete sou yon ne epi imaj li ponpe soti pou yon tan trè long, Lè sa a, yon lòt gous ki "frape" menm ne pral tou senpleman. pa kòmanse rale imaj la nan gous la nouvo. Olye de sa, li tann jiskaske imaj la nan gous anvan an rale. Kòm yon rezilta, yon gous ki te deja pwograme epi ki gen imaj ki te kapab telechaje nan jis yon minit pral fini nan estati a nan. containerCreating.
Evènman yo pral sanble yon bagay tankou sa a:
Normal Pulling 8m kubelet, ip-10-241-44-128.ap-northeast-1.compute.internal pulling image "registry.example.com/infra/openvpn/openvpn:master"Li sanble ke yon sèl imaj ki soti nan yon rejis dousman ka bloke deplwaman pou chak ne.
Malerezman, pa gen anpil fason pou soti nan sitiyasyon an:
- Eseye itilize Rejis Docker ou dirèkteman nan gwoup la oswa dirèkteman ak gwoup la (pa egzanp, GitLab Rejis, Nexus, elatriye);
- Sèvi ak sèvis piblik tankou .
Istwa 5. Nœuds pann akòz mank de memwa
Pandan operasyon an nan aplikasyon divès kalite, nou te rankontre tou yon sitiyasyon kote yon ne konplètman sispann aksesib: SSH pa reponn, tout demon siveyans tonbe, ak Lè sa a, pa gen anyen (oswa prèske pa gen anyen) anomali nan mòso bwa yo.
Mwen pral di w nan foto lè l sèvi avèk egzanp yon sèl ne kote MongoDB te fonksyone.
Sa a se sa anlè a sanble до aksidan:
Epi tankou sa a - apre aksidan:
Nan siveyans, gen tou yon so byen file, nan ki ne a sispann disponib:
Kidonk, nan Ekran yo li klè ke:
- RAM sou machin nan tou pre fen;
- Gen yon so byen file nan konsomasyon RAM, apre yo fin ki aksè a tout machin nan se sibitman enfim;
- Yon gwo travay rive sou Mongo, ki fòse pwosesis DBMS yo sèvi ak plis memwa ak li aktivman nan disk.
Li sanble ke si Linux kouri soti nan memwa gratis (presyon memwa mete nan) epi pa gen okenn swap, Lè sa a до Lè asasen an OOM rive, yon aksyon balanse ka leve ant voye paj nan kachèt paj la epi ekri yo tounen sou disk. Sa a se fè pa kswapd, ki kouraz libere paj memwa anpil ke posib pou distribisyon ki vin apre.
Malerezman, ak yon gwo chaj I/O makonnen ak yon ti kantite memwa gratis, kswapd vin anbouteyaj nan tout sistèm nan, paske yo mare ak li tout alokasyon (paj defo) nan paj memwa nan sistèm nan. Sa a ka kontinye pou yon tan trè lontan si pwosesis yo pa vle sèvi ak memwa ankò, men yo fiks nan kwen an menm nan gwo twou san fon an OOM-asasen.
Kesyon natirèl la se: poukisa asasen an OOM vini an reta? Nan iterasyon aktyèl li yo, asasen an OOM trè estipid: li pral touye pwosesis la sèlman lè tantativ pou asiyen yon paj memwa echwe, sa vle di. si fòt paj la echwe. Sa a pa rive pou yon bon bout tan, paske kswapd kouraz libere paj memwa, jete kachèt paj la (tout disk I/O nan sistèm nan, an reyalite) tounen nan disk. Nan plis detay, ak yon deskripsyon etap ki nesesè yo elimine pwoblèm sa yo nan nwayo a, ou ka li .
Konpòtman sa a ak Linux Kernel 4.6+.
Istwa 6. Gous yo bloke nan eta annatant
Nan kèk grap, kote gen reyèlman anpil gous opere, nou te kòmanse remake ke pi fò nan yo "kwoke" pou yon tan trè long nan eta a. Pending, byenke resipyan yo Docker tèt yo deja kouri sou nœuds yo epi yo ka travay avèk manyèlman.
Anplis, nan describe pa gen anyen ki mal:
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 containerApre kèk fouye, nou te fè sipozisyon ke kubelet la tou senpleman pa gen tan voye tout enfòmasyon sou eta a nan gous yo ak tès vivan / preparasyon pou sèvè API a.
Apre yo fin etidye èd, nou jwenn paramèt sa yo:
--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)Jan yo wè, valè default yo byen piti, ak nan 90% yo kouvri tout bezwen ... Sepandan, nan ka nou an sa a pa t 'ase. Se poutèt sa, nou mete valè sa yo:
--event-qps=30 --event-burst=40 --kube-api-burst=40 --kube-api-qps=30 --registry-qps=30 --registry-burst=40... ak rekòmanse kubelets yo, apre sa nou te wè foto sa a nan graf yo nan apèl nan sèvè API a:
... e wi, tout bagay te kòmanse vole!
PS
Pou èd yo nan kolekte pinèz ak prepare atik sa a, mwen eksprime gwo rekonesans mwen an bay anpil enjenyè nan konpayi nou an, epi espesyalman nan kòlèg mwen nan ekip R&D nou an Andrey Klimentyev ().
P
Li tou sou blog nou an:
- «'.
- Kubernetes konsèy ak ke trik nouvèl bouk:
- «";
- «";
- «";
- «'.
Sous: www.habr.com