text stringlengths 0 1.8k |
|---|
[2118.76 --> 2120.56] fragments that you can test |
[2120.56 --> 2122.56] locally so i believe in the |
[2122.56 --> 2124.52] ideas that you should have as |
[2124.52 --> 2126.60] little logic as possible in |
[2126.60 --> 2128.22] your ci proprietary |
[2128.22 --> 2130.08] orchestration language on that |
[2130.08 --> 2131.86] you should group these in |
[2131.86 --> 2133.76] typically make files to help |
[2133.76 --> 2135.54] the stability of the system |
[2135.54 --> 2136.98] oleg firstly i agree that you |
[2136.98 --> 2137.80] should be able to test |
[2137.80 --> 2139.14] locally but it doesn't mean |
[2139.14 --> 2140.98] that you cannot use pipeline |
[2140.98 --> 2142.48] definitions because many |
[2142.48 --> 2144.42] modern systems actually allow |
[2144.42 --> 2146.12] running pipelines locally it's |
[2146.12 --> 2147.40] not just jenkins so for |
[2147.40 --> 2148.52] jenkins we have jenkins file |
[2148.52 --> 2150.34] runner for team city you can |
[2150.34 --> 2152.30] run conclin dsl for github |
[2152.30 --> 2154.78] there are projects as well and |
[2154.78 --> 2157.42] it basically poses this gap so |
[2157.42 --> 2158.54] if you have proper |
[2158.54 --> 2160.10] configuration management for a |
[2160.10 --> 2161.86] system if you can produce your |
[2161.86 --> 2163.24] production cd environment |
[2163.24 --> 2165.18] locally for example if you run |
[2165.18 --> 2167.74] the system in the container you |
[2167.74 --> 2169.50] can easily do local development |
[2169.50 --> 2170.98] and create corporate x pipelines |
[2170.98 --> 2172.56] so that's a good solution we |
[2172.56 --> 2173.64] will talk about pipeline |
[2173.64 --> 2175.04] development what it looks like |
[2175.04 --> 2177.16] but i would like to go back to |
[2177.16 --> 2179.44] the production question how would |
[2179.44 --> 2180.56] you deploy jenkins in |
[2180.56 --> 2182.24] production so i think cyril was |
[2182.24 --> 2183.88] mentioning kubernetes you |
[2183.88 --> 2185.46] would deploy jenkins a |
[2185.46 --> 2186.62] production deployment and you |
[2186.62 --> 2187.68] would manage jenkins via |
[2187.68 --> 2189.56] kubernetes and i imagine helm |
[2189.56 --> 2191.20] chart or operator what would you |
[2191.20 --> 2193.16] go cyril which way i am not |
[2193.16 --> 2194.88] knowledgeable enough okay what |
[2194.88 --> 2196.54] about you like i would go with |
[2196.54 --> 2199.08] helm chart to be honest because |
[2199.08 --> 2200.70] helm chart allows to be more |
[2200.70 --> 2202.10] flexible in terms of defining |
[2202.10 --> 2204.78] system okay operator has a lot of |
[2204.78 --> 2206.38] advantages if you want to build a |
[2206.38 --> 2207.98] reactive system so which is |
[2207.98 --> 2209.14] basically based on kubernetes |
[2209.14 --> 2210.08] apis |
[2210.08 --> 2212.04] text to some events automatically |
[2212.04 --> 2215.24] scales etc but for jenkins to my |
[2215.24 --> 2217.20] experience it's not always needed |
[2217.20 --> 2219.52] it can be used in particular use |
[2219.52 --> 2219.86] cases |
[2219.86 --> 2223.38] so i would go with operators only if i |
[2223.38 --> 2226.18] was building highly available jenkins |
[2226.18 --> 2227.98] solution where i would be managing |
[2227.98 --> 2229.76] controllers automatically |
[2229.76 --> 2232.42] provisioning them and if i had to |
[2232.42 --> 2234.28] share the context between them |
[2234.28 --> 2235.98] okay right now it's not quite |
[2235.98 --> 2238.50] possible with stock jenkins so i |
[2238.50 --> 2240.10] would rather go after helm chart |
[2240.10 --> 2242.32] in that world where you deploy a |
[2242.32 --> 2243.28] where you have a production |
[2243.28 --> 2246.50] deployment of jenkins using helm how |
[2246.50 --> 2248.92] would you configure the pipelines how |
[2248.92 --> 2250.10] do you configure jenkins and then how |
[2250.10 --> 2251.36] would you configure for example the |
[2251.36 --> 2253.26] agents themselves where would that |
[2253.26 --> 2255.26] happen how that look like everything |
[2255.26 --> 2255.98] is code |
[2255.98 --> 2257.90] okay because currently if you talk |
[2257.90 --> 2259.58] about pipelines if you use jenkins |
[2259.58 --> 2261.86] pipeline drop this all these |
[2261.86 --> 2264.12] technologies can be stored as code in |
[2264.12 --> 2264.86] your repository |
[2264.86 --> 2267.26] in parallel with your project so that |
[2267.26 --> 2268.96] when you build your project you have a |
[2268.96 --> 2270.98] pipeline and you can test them all |
[2270.98 --> 2271.24] together |
[2271.24 --> 2273.46] and basically the same for agent |
[2273.46 --> 2274.02] definitions |
[2274.02 --> 2275.88] okay for example if you use kubernetes |
[2275.88 --> 2278.06] plugin you can store agent definition |
[2278.06 --> 2280.52] again in the same repository so that |
[2280.52 --> 2282.64] you have your build system within your |
[2282.64 --> 2284.86] project and it's portable or you can |
[2284.86 --> 2286.94] have it separately if needed but still |
[2286.94 --> 2288.78] it should be defined as code somewhere |
[2288.78 --> 2290.96] and i would argue that actually the |
[2290.96 --> 2293.20] entire combination of jenkins so for |
[2293.20 --> 2295.32] us it's a server itself login |
[2295.32 --> 2297.88] configuration pipeline libraries you use |
[2297.88 --> 2300.62] and the default pipeline building blocks |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.