)]}'
{"doc/source/install/refarch/common.rst":[{"author":{"_account_id":11655,"name":"Julia Kreger","email":"juliaashleykreger@gmail.com","username":"jkreger","status":"Flying to the moon with a Jetpack!"},"change_message_id":"2ca3e724c2f849b3893f26ffe81c90a386dc99b0","unresolved":true,"context_lines":[{"line_number":115,"context_line":"Otherwise, it is recommended to use UEFI HTTP boot or iPXE, when it is"},{"line_number":116,"context_line":"supported by target hardware. The BMC network is not an issue in this case, but"},{"line_number":117,"context_line":"the requirement of DHCP on a shared L2 network will limit the number of"},{"line_number":118,"context_line":"simultaneously running processes to maximum a few dozens."},{"line_number":119,"context_line":""},{"line_number":120,"context_line":"Hardware specifications"},{"line_number":121,"context_line":"~~~~~~~~~~~~~~~~~~~~~~~"}],"source_content_type":"text/x-rst","patch_set":2,"id":"588355d2_478616aa","line":118,"range":{"start_line":118,"start_character":15,"end_line":118,"end_character":57},"updated":"2026-07-22 14:11:01.000000000","message":"Per what, subnet?","commit_id":"7854b5a42b7b5919652c03976a5605ae3f3ff34c"},{"author":{"_account_id":11655,"name":"Julia Kreger","email":"juliaashleykreger@gmail.com","username":"jkreger","status":"Flying to the moon with a Jetpack!"},"change_message_id":"2ca3e724c2f849b3893f26ffe81c90a386dc99b0","unresolved":true,"context_lines":[{"line_number":307,"context_line":"reliability and performance. This is a very safe bet across different"},{"line_number":308,"context_line":"configurations and drivers.  There is some tolerance for a larger number per"},{"line_number":309,"context_line":"conductor. However, it was reported [1]_ [2]_ that reliability may degrade when"},{"line_number":310,"context_line":"handling approximately 300 bare metal nodes per conductor.  To handle a much"},{"line_number":311,"context_line":"higher number of machines, you need to scale your cluster horizontally, and/or"},{"line_number":312,"context_line":"apply performance tuning."},{"line_number":313,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"366ca117_66a6f4dd","line":310,"updated":"2026-07-22 14:11:01.000000000","message":"I will note, I\u0027ve talked with folks doing 800-1000 *per* conductor today with relatively stock defaults without any issues.  If we have data to say 300, lets go with it, but we may want to move that upwards a little. Also, that 300 measurement was when we still ran with eventlet if memory serves.","commit_id":"7854b5a42b7b5919652c03976a5605ae3f3ff34c"},{"author":{"_account_id":11655,"name":"Julia Kreger","email":"juliaashleykreger@gmail.com","username":"jkreger","status":"Flying to the moon with a Jetpack!"},"change_message_id":"2ca3e724c2f849b3893f26ffe81c90a386dc99b0","unresolved":true,"context_lines":[{"line_number":317,"context_line":""},{"line_number":318,"context_line":"To get a higher nodes-per-conductor ratio, consider these tricks:"},{"line_number":319,"context_line":""},{"line_number":320,"context_line":"- Prefer Redfish or other HTTP-based drivers to using IPMI."},{"line_number":321,"context_line":"- Prefer virtual media boot over a performant BMC network to using network boot"},{"line_number":322,"context_line":"  (see `boot interface`_)."},{"line_number":323,"context_line":"- Serve raw images directly from Swift or from an external HTTP server using"}],"source_content_type":"text/x-rst","patch_set":2,"id":"b2e0ed33_6f370326","line":320,"updated":"2026-07-22 14:11:01.000000000","message":"Is there any way we can answer the question of \"why?\" i.e. why use redfish over ipmi?","commit_id":"7854b5a42b7b5919652c03976a5605ae3f3ff34c"},{"author":{"_account_id":11655,"name":"Julia Kreger","email":"juliaashleykreger@gmail.com","username":"jkreger","status":"Flying to the moon with a Jetpack!"},"change_message_id":"2ca3e724c2f849b3893f26ffe81c90a386dc99b0","unresolved":true,"context_lines":[{"line_number":327,"context_line":"- Significantly decrease frequency of various periodic processes, including the"},{"line_number":328,"context_line":"  :doc:`power synchronization loop \u003c/admin/power-sync\u003e`. Turn off periodic"},{"line_number":329,"context_line":"  tasks for feature you don\u0027t use (e.g. firmware upgrades, BIOS settings)."},{"line_number":330,"context_line":"- If you don\u0027t use Nova, detach provisioned nodes from Ironic by deleting them"},{"line_number":331,"context_line":"  in maintenance mode. :doc:`Adopt \u003c/admin/adoption\u003e` them again for further"},{"line_number":332,"context_line":"  management."},{"line_number":333,"context_line":"- If operating an OpenStack cloud with a busy messaging bus, check if using"},{"line_number":334,"context_line":"  JSON RPC makes the API more responsive. For a single-conductor Ironic,"},{"line_number":335,"context_line":"  compare the all-in-one ``ironic`` script against using WSGI and RPC."}],"source_content_type":"text/x-rst","patch_set":2,"id":"4e7a10d0_505b4e7c","line":332,"range":{"start_line":330,"start_character":1,"end_line":332,"end_character":13},"updated":"2026-07-22 14:11:01.000000000","message":"This one gives me a little heartburn. Perhaps we could suggest \"if management actions won\u0027t be needed for some time...\" ?","commit_id":"7854b5a42b7b5919652c03976a5605ae3f3ff34c"},{"author":{"_account_id":35929,"name":"Mahnoor Asghar","display_name":"Mahnoor Asghar","email":"masghar@redhat.com","username":"mahnoorasghar"},"change_message_id":"eda09258bbbf022d790cc40ec2e92dd5fbed6384","unresolved":true,"context_lines":[{"line_number":327,"context_line":"- Significantly decrease frequency of various periodic processes, including the"},{"line_number":328,"context_line":"  :doc:`power synchronization loop \u003c/admin/power-sync\u003e`. Turn off periodic"},{"line_number":329,"context_line":"  tasks for feature you don\u0027t use (e.g. firmware upgrades, BIOS settings)."},{"line_number":330,"context_line":"- If you don\u0027t use Nova, detach provisioned nodes from Ironic by deleting them"},{"line_number":331,"context_line":"  in maintenance mode. :doc:`Adopt \u003c/admin/adoption\u003e` them again for further"},{"line_number":332,"context_line":"  management."},{"line_number":333,"context_line":"- If operating an OpenStack cloud with a busy messaging bus, check if using"},{"line_number":334,"context_line":"  JSON RPC makes the API more responsive. For a single-conductor Ironic,"},{"line_number":335,"context_line":"  compare the all-in-one ``ironic`` script against using WSGI and RPC."}],"source_content_type":"text/x-rst","patch_set":2,"id":"b1c22f98_2ef1fea0","line":332,"range":{"start_line":330,"start_character":1,"end_line":332,"end_character":13},"in_reply_to":"4e7a10d0_505b4e7c","updated":"2026-07-22 16:30:46.000000000","message":"++, I was a little surprised here as well","commit_id":"7854b5a42b7b5919652c03976a5605ae3f3ff34c"}]}