viplab tcp 31.08.2026 — Transcript

Full transcript

1. 0:06Hi Max. Hi Christian.
2. 0:10>> Hi.
3. 0:11>> How are you? Everything's good.
4. 0:57Hi Stefan. Hi Yan.
5. 1:01>> Yeah.
6. 1:03>> Okay. Let me check with TK if he's able
7. 1:06to join.
8. 1:47Hi all. Sorry for being late.
9. 1:48>> Hi TK. No problem.
10. 1:50>> Yeah.
11. 1:51>> Okay. So I think we can start. Stefan,
12. 1:53do we need to wait? Anyone else from the
13. 1:56VL Vlab? Sorry besides
14. 2:00>> No. Okay.
15. 2:00>> No, we can start.
16. 2:03>> Okay. So for from what I understood for
17. 2:06the ES SV2 so the ESXi
18. 2:10upgrade so this depends uh first on the
19. 2:13validating the uh firmware strategy
20. 2:17right so Utar uh mentioned during uh the
21. 2:22last weeks I think uh that you
22. 2:24identified the SSP uh 201241.02
23. 2:2902 with uh the one view uh 11.3 as the
24. 2:35recommended transition and combination
25. 2:38uh that will support both ESXi 7 and um
26. 2:428.
27. 2:44Um
28. 2:47so we need to validate u this approach
29. 2:49in vlab as well and to
30. 2:53confirm the upgrade path and u yeah the
31. 2:58operational approach and the next steps.
32. 3:02>> Mhm. Uh and uh Max you mentioned that uh
33. 3:08the version um uh the SSP version um
34. 3:13identified by TK and recommended
35. 3:17uh will be somehow uh
36. 3:21um
37. 3:23will expire soon.
38. 3:26So um
39. 3:29I'm not sure if um so this is already
40. 3:32approaching approaching to end of
41. 3:34support from what I understood.
42. 3:39Yeah, I think it's even already end of
43. 3:42support uh as HP gives I think 18 months
44. 3:46uh support on on an S SP or SP.
45. 3:51Um and as it is from end of 2024,
46. 3:56these are already over.
47. 3:58Um but yeah as as long as we have to
48. 4:03keep ESXi7
49. 4:05uh we cannot go higher because of HP
50. 4:08support.
51. 4:10Um so
52. 4:13anyway we we are locked on on this
53. 4:16version
54. 4:19>> but uh if you want if you want to
55. 4:21upgrade to ESXi8
56. 4:24as TK proposed the SSP uh 2024.11
57. 4:30can we use this one?
58. 4:33Yeah, that's the only one which
59. 4:35>> because it's the one only one comp
60. 4:36compatible the
61. 4:38with
62. 4:39>> it's the only one which supports ESXi 7
63. 4:42and 8. Yeah.
64. 4:43>> And um I think in in
65. 4:48>> at least part of of uh the ES blocks uh
66. 4:52we have to stay on on seven on uh
67. 4:56specific clusters. So uh yeah we cannot
68. 4:59go higher with the SSP.
69. 5:04>> Ahu okay that's uh I I haven't got the
70. 5:08clue why we cannot update but you say
71. 5:10there are some projects which uh want to
72. 5:14stay on ASX7 on the EIS.
73. 5:18>> Um I'm not sure if they want but uh
74. 5:22>> okay they have to because of I don't
75. 5:24know they are not compatible. Okay,
76. 5:27whatever.
77. 5:28>> Yeah,
78. 5:29>> let's say we can we cannot uh upgrade
79. 5:32the whole environment uh relatively
80. 5:36quickly due to incompatibilities
81. 5:39of some of the VNFs with E68.
82. 5:42So we need to wait until they in some
83. 5:46cases until they complete the migration
84. 5:49and the the decommissioning or in other
85. 5:52cases until they uh become compatible uh
86. 5:57once they upgrade their modes.
87. 5:59>> Yeah. So
88. 6:00>> I haven't Yeah. So which on which SSP
89. 6:03version is now compatible with both
90. 6:06versions? I haven't get this
91. 6:09>> 2041.
92. 6:11Yeah, that is
93. 6:15>> 2024.11
94. 6:17is the uh compatible version for both
95. 6:20ESXi 7 and 8.
96. 6:22>> Okay. Okay. Yeah, I see. Mhm.
97. 6:25>> What we do currently have uh in
98. 6:28production t do you know what version do
99. 6:32we have?
100. 6:33>> One.
101. 6:34So one one version earlier. Yeah.
102. 6:37>> Okay. Okay.
103. 6:39and is the same in the vlo right
104. 6:43>> that's a question to Jan and Christian
105. 6:45can you can you check it in which
106. 6:48version we are would be also different
107. 6:51between TS4 and TS5
108. 6:56>> so I think in lab it depends from TS4 to
109. 7:01TS5 and inside TS5 I think uh from one
110. 7:06it's different from one cluster to
111. 7:08another
112. 7:08>> Mhm. M also. Yes.
113. 7:10>> But the good thing is that in in in
114. 7:14TS5 we have uh sorry the version
115. 7:19uh 24.11
116. 7:22I think installed.
117. 7:25So we can test basically this uh this
118. 7:29path which is shown here.
119. 7:36So, we can test in the TS5, right?
120. 7:39Because uh we have a
121. 7:42>> Yes,
122. 7:44I'm not sure if it's still interest.
123. 7:50>> I think
124. 7:53>> what is uh let me try to find the email.
125. 8:00What is important to test I think is the
126. 8:03combination of SSP 20 2024.11
127. 8:08with uh one view 11.3
128. 8:12according to HP site it is supported uh
129. 8:18>> but we need to we still need to test it
130. 8:22>> but I think still we we first can only
131. 8:27go to uh I think 10 version 10.
132. 8:32Uh that is the latest one still
133. 8:36supporting 202401.
134. 8:40So the actual installed one.
135. 8:43Then we would have to do all the
136. 8:45firmware updates. Um afterwards we can
137. 8:49go to uh one view 11. Right.
138. 8:55>> Right. Yes. I think we need to even go
139. 8:57to one one view 9.2. to I need to double
140. 9:01check again.
141. 9:02>> Yeah, I think 10.1 or something was the
142. 9:05latest one supporting uh 202401.
143. 9:10>> Okay, so they have updated their site
144. 9:13then
145. 9:13>> I'm I'm not sure.
146. 9:15>> I would have to check again.
147. 9:17>> Yeah,
148. 9:18>> already a few days ago. But I did so
149. 9:23>> it's more more recent than let me check.
150. 9:26I think I wrote it down.
151. 9:30So I'm just logged into PS5 years and
152. 9:33one is on 11.10
153. 9:37>> currently
154. 9:39>> 11.10.
155. 9:40>> Yeah.
156. 9:43>> All sides or
157. 9:46PS5? Yes.
158. 9:47>> ES5. Okay. Should I check PS4?
159. 9:54M yes would be nice.
160. 9:58>> I wrote it down. 10.10 is the latest
161. 10:01with uh support for 2020 41.
162. 10:11>> So the idea is to go from uh one view
163. 10:158.7 to
164. 10:20matter if we go to 9.2 or 10.10. Um
165. 10:27anyway, we we would have to do the next
166. 10:30steps afterwards um after the the tender
167. 10:34update.
168. 10:37So it doesn't really matter if we wait
169. 10:38on on 10 uh 10 or on 9.2.
170. 10:43>> In TS4 years, we on 11.01.
171. 10:53So both on version 11.
172. 10:56>> Yeah.
173. 10:57>> Okay.
174. 11:00Uh
175. 11:02so
176. 11:12but you I think you don't know how the
177. 11:15update was done in in the past, right?
178. 11:19if it was done in depth or
179. 11:23directly or
180. 11:29no not anymore.
181. 11:33>> But I see when I look at the um
182. 11:37SSP versions, we have mainly 2024.11.
183. 11:43Okay, some are all also higher. Maybe
184. 11:46there was some test
185. 11:49I saw something with also 2025.
186. 11:54>> Yeah, a few are on 2025. They're mostly
187. 11:56on 2024.
188. 11:59Yeah, I mean the the actual one would be
189. 12:0120264.
190. 12:04I think that would be the latest one.
191. 12:08>> I think there is a 07
192. 12:10>> 07 now even. Yeah, but of course that
193. 12:14would only be compatible with ESXi8.
194. 12:18>> Yeah.
195. 12:21>> So I think in general it would a mixed
196. 12:25approach. Um I think we we have a few
197. 12:29sides where we can update everything on
198. 12:32eight, right? And a few sides where we
199. 12:35have some clusters on seven.
200. 12:40Okay. And TS4. Yes. I see we have uh the
201. 12:44the SSDs already on 20264.
202. 12:55Yeah, but it it depends. So service
203. 12:59clusters have not been updated
204. 13:04in TS in TS5.
205. 13:07Uh so on the service clusters we still
206. 13:11have 2024.11.
207. 13:20So
208. 13:22these sequence that I see is first you
209. 13:26upgrade one view from 8 to 10 upgrade
210. 13:30SSP to 20 24
211. 13:34uh.2.11 2 to.1
212. 13:38upgrading one view to uh the latest
213. 13:42version well probably maybe to 11.3
214. 13:47because the latest uh I'm not sure will
215. 13:50would support 2024.11
216. 13:53and then depending on the progress uh as
217. 13:58soon as we get some enclosures
218. 14:01that are um let's I fully upgraded to
219. 14:06ESXi8. We can progress with 20 uh with
220. 14:10the upgrade from 2024
221. 14:13uh.11 to 2026.
222. 14:16I mean it it it depends a bit. Um
223. 14:21I think on on the sides or we we may
224. 14:26have two different target uh SSPs for
225. 14:32the sites where we have to stay on on
226. 14:34ESX i7. In some clusters we we have
227. 14:38202411 as target SSP.
228. 14:43And for the sites where we can update
229. 14:47all clusters on ESXi8, I think we may be
230. 14:51able to go directly to the latest SSP.
231. 14:58>> So for some sites, we don't need to do
232. 15:01multiple jumps and for the other ones we
233. 15:03will need to
234. 15:05go first from one version to another and
235. 15:09then from the SSP. Yeah. Right. So
236. 15:12that's what you want to say?
237. 15:16>> I think so. Um and and just thinking
238. 15:20about I mean for the
239. 15:25chassis themselves,
240. 15:29we may have the situation that we have
241. 15:33to update the chassis of course before
242. 15:36the blade firmware.
243. 15:41Uh,
244. 15:43and I mean that that would be a
245. 15:45situation not fully supported, right?
246. 15:48when we go with the chassis firmware to
247. 15:52the latest SSP
248. 15:54and have the blade firmware still on on
249. 15:57the old 2024
250. 16:00uh 01.
251. 16:09I think is not too many things are
252. 16:12supported. Yeah,
253. 16:14>> officially supported out.
254. 16:19And I'm not sure if we would be able to
255. 16:22do the updates chassis by chassis.
256. 16:26So do the chassis and all the hosts at
257. 16:29one step.
258. 16:33That would be the easiest way and the
259. 16:36safest way of course.
260. 16:40Um, but I'm not sure if we have the
261. 16:42resources in in the clusters.
262. 16:47>> Yeah, because we need to offload 12.
263. 16:52>> Yeah.
264. 16:57>> I don't know.
265. 17:04No for for production that's
266. 17:09be possible.
267. 17:12>> I think hasn't been invited. I mean he
268. 17:17can't tell us how exactly he did.
269. 17:20>> I think so but I think also his
270. 17:23indication today.
271. 17:25>> Okay.
272. 17:27>> So we
273. 17:31Sorry I'm muted. So this explains a lot.
274. 17:33Um, this is this is his first day off.
275. 17:38Oh, so he's back again this week. Oh,
276. 17:41yes.
277. 17:45>> Oh, I'm sorry. I was not aware that he
278. 17:48he should be invited as well.
279. 17:51>> No, not
280. 17:52a problem. But but what is the question
281. 17:55to to Army then if he have done this
282. 17:57with one step or with
283. 18:00several steps?
284. 18:03I mean, for example, as Mark said, uh
285. 18:07can we upgrade the chassis first and
286. 18:10then the blades one by one? Probably
287. 18:14uh over uh several nights
288. 18:18>> in in general that possible. We we did
289. 18:20it
290. 18:21>> yes
291. 18:21>> in the past but I'm not sure if that
292. 18:24would be possible for a direct step from
293. 18:27from 2024
294. 18:29uh 01 to 2026 07 for example
295. 18:37cuz I mean in the meantime you would
296. 18:40have the latest firmware on the chassis
297. 18:44components like VCF
298. 18:46and you will have the very old firmware
299. 18:49on the blade
300. 18:53and that combination
301. 18:56I think it will not be supported from
302. 18:58HPE and we don't know if it's working
303. 19:05>> right
304. 19:13but you're meaning this in production
305. 19:15but in the V vlab can We should should
306. 19:19we perform multiple jumps or can we go
307. 19:23from one version to the latest version
308. 19:26directly?
309. 19:27>> I mean we already did some updates. Um
310. 19:30the question would be if if we can test
311. 19:32this but
312. 19:34>> even if we test this
313. 19:36we only see if it's in general working
314. 19:39but we will not see if if there are
315. 19:42specific problems right.
316. 20:08So sorry.
317. 20:11>> I think mainly the the thing is to
318. 20:14investigate
319. 20:16uh if you would be able to do the update
320. 20:19by
321. 20:26>> so we need this confirmation from air
322. 20:28you mentioned yeah
323. 20:32>> no I mean I'm not sure who can
324. 20:35investigate on on this depends on
325. 20:38capacity in the clusters if it would be
326. 20:41able to maybe down VMs and and such
327. 20:46things.
328. 21:02I mean for sure we don't have enough
329. 21:04space to offload 12 blades
330. 21:11in in most of the uh clusters.
331. 21:15to not sell all of them
332. 21:17>> at least if we do it in in the
333. 21:22maybe call it old old fashioned way. So
334. 21:25uh do the update
335. 21:28uh host yes and and under running
336. 21:31applications.
337. 21:32>> Yeah.
338. 21:34>> Um
339. 21:34>> I mean that's the only way that's how we
340. 21:38get it for veno for example.
341. 21:41I would not say that's the only way that
342. 21:44is
343. 21:46the
344. 21:47yeah maybe maybe the easiest way from
345. 21:51from our view because we don't have to
346. 21:54to ask applications to shut down uh VMs
347. 21:59and they don't have to be involved but
348. 22:03um
349. 22:04on the other hand
350. 22:06yeah I mean
351. 22:10The faster way would be to offload uh
352. 22:13VNFS and and shut them down and do a
353. 22:20full cluster in one night for example
354. 22:22with switched off again and do it
355. 22:25chassis by chassis as we did it in in
356. 22:29cars.