many systemd services for manageiq

[Jonathan-Caruana]

Hello,

I wondering if it's normal to have many many systemd services for manageIQ? Some of them are just the same service but with an different ID.

We currently have 1060 systemd services on our MiQ instance, for example for manageiq-web_service :

manageiq-web_service@03090a39-d92d-41d5-a4b4-1fff8e30bb70.service
manageiq-web_service@0339fc09-1f2f-44c1-8bac-a780abd9c7f0.service
manageiq-web_service@0a2aab59-7a2f-4f5c-9666-73bbae032818.service
manageiq-web_service@0bcfd0c4-e4be-4169-9bdd-6cd2d0079178.service
manageiq-web_service@11b08c48-1669-4c46-852e-4de370940eb4.service
manageiq-web_service@214020e7-09a8-4dfd-8f15-4f704fb1cae3.service
manageiq-web_service@21e214a8-f498-4777-8b59-2ab93c8b1cb2.service
manageiq-web_service@225fd293-79c8-44e0-93d0-af8e93dd46a7.service
manageiq-web_service@25c6930e-70ff-435f-a081-b15f1df9e87d.service
manageiq-web_service@262485c5-9171-4484-bb8d-32eb46ff5a66.service
manageiq-web_service@26c36d3e-91ca-4ad4-9a78-2b58216e55ab.service
manageiq-web_service@293d7684-f6b1-4c30-9a2e-3d2208fc4839.service
manageiq-web_service@2d3f14c9-0347-4997-96ca-0f73c5a51945.service
manageiq-web_service@30a096cb-bd4e-4246-8ece-eda9eca6d321.service
manageiq-web_service@342bf2b9-2c36-4ba5-a5a5-94e7363a8278.service
manageiq-web_service@350fff78-565e-4834-9e7a-7a3b8e154132.service
manageiq-web_service@3a32f344-0dcf-48ca-b4f2-ec44c132b126.service
manageiq-web_service@3c30dce8-ab04-447f-a0da-2d54df381d81.service
manageiq-web_service@420f1a57-98cb-4a42-9c25-1d328185cadd.service
manageiq-web_service@42be2a07-28e5-42c3-963d-c86ad20eebe3.service
manageiq-web_service@477d7ff6-b14e-4ec7-8b1d-a1550d4aea91.service
manageiq-web_service@4d2003dc-ad64-405e-98be-0df5474819ba.service
manageiq-web_service@4e741bb7-e6f9-4698-8d98-8841b28ec3eb.service
manageiq-web_service@4e7b1dca-df13-4a33-9df1-0cfc89fb0b59.service
manageiq-web_service@58ede861-e7ba-458a-acd7-0acd03db3a80.service
manageiq-web_service@5e9f486f-af2c-4835-a46d-0fc4a6aa8ee3.service
manageiq-web_service@60a94e38-112a-408a-8190-03d5ab3ed37d.service
manageiq-web_service@68126a18-9142-4d92-bd66-63a3c80cc6ae.service
manageiq-web_service@73a67bef-0f41-43f3-b060-fc868bc04d9a.service
manageiq-web_service@7f18e3a3-2a1d-4a1a-9542-5618016cf9dd.service
manageiq-web_service@894decbc-dc00-41bd-8dd9-be9129476cac.service
manageiq-web_service@8d8d7a34-a3a0-4a8c-b6fd-0513fd3558a1.service
manageiq-web_service@9aef892a-436d-481f-84e0-c00c888abce0.service
manageiq-web_service@9f0e9899-173f-413b-931a-93be1c530849.service
manageiq-web_service@a1140de3-c895-4569-98cc-6fb5cfe23fe6.service
manageiq-web_service@a91fc7eb-44c2-4d0d-880e-23f78becbc0f.service
manageiq-web_service@afeffa02-d1b4-4593-a12b-60bbce9ea9b6.service
manageiq-web_service@bb4340c0-d48f-40de-bb92-0930a1cfa73e.service
manageiq-web_service@c5c263ec-1054-47c4-9619-83f04cbf98e1.service
manageiq-web_service@c79fa2eb-0f5b-42fc-b8b0-0349524a1ad5.service
manageiq-web_service@cdbb64ab-20d7-40c7-a264-3d71458c1de5.service
manageiq-web_service@cfa3437d-127d-4c3c-907f-f154188ca9e9.service
manageiq-web_service@dd1c54d6-7810-42d7-a843-c874d01dd1c4.service
manageiq-web_service@df887f94-6e0e-4db5-b9c2-4af66c848719.service
manageiq-web_service@e82cbacf-452a-48e8-a1ad-feb2132f87f8.service
manageiq-web_service@e9b388ce-bed6-4e91-b977-1aab1b868bdd.service
manageiq-web_service@ea41fabc-c6d2-43ac-8f23-55b1c19f9e80.service
manageiq-web_service@ea91ccf8-319c-44e2-a29c-8b407b5581a6.service
manageiq-web_service@ec87f4dc-c716-4b17-a446-8b4af39ec0ce.service
manageiq-web_service@ee493e28-e3a4-4ff1-b5c6-ef0256d527a8.service
manageiq-web_service@f07c255b-452c-41df-a1e4-88f8052d3e59.service
manageiq-web_service@fbb7f054-7a1d-427e-81bf-ee80bf283605.service

Most of them are inactive (dead), so i'm curious to know if it's safe to remove them?

Maybe it's because of successive upgrade.

Regards,

[agrare]

Hey @Jonathan-Caruana these should be cleaned up by evmserverd but we might be missing some case where they aren't failed but also aren't active.
If you systemctl stop evmserverd then yes it is safe to cleanup all manageiq-*@*.service instance units.

[kbrock]

Hello @Jonathan-Caruana - yes, this is not normal. The evmserverd process should delete the services.
The valid services are in the miq_workers table.

select system_uid from miq_workers;

OR

MiqWorker.pluck(:guid)

UPDATE: As @agrare mentioned, if you stop evmserverd, there should be no values in miq_workers (or at least all values in there are no longer valid) and no manageiq-*@*.service instance units.

[Jonathan-Caruana]

Thank you @agrare and @kbrock for your answers, i will be able to cleaned up :)