In order to troubleshot and experiment with asw2-a-eqiad for T201145, we need to move servers off of that stack.
This represents 17 servers to move, of which 7 will need to either:
- Be moved to a different rack
- Have a cross rack uplink
- Be connected to an extra 10G switch in the same rack
My guess is that cross rack uplink is the easiest temporary solution.
If some servers are not production (and can wait before becoming so) they could be left on asw2-a to serve as canaries.
| Description | asw2 port | Move to | Special care | Notes |
|---|---|---|---|---|
| dns1001 | ge-1/0/8 | asw-a:ge-1/0/8 | Yes | |
| lvs1015:enp5s0f0 | xe-2/0/0 | asw2-a5(ex4500):xe-0/0/2 | No | Will utilize current cross connect - downtime may be up to 1 min |
| cloudelastic1001 | xe-2/0/7 | asw2-a5(ex4500):xe-0/0/3 | No | 10M Fiber #3935 |
| lvs1016:enp4s0f1 {#3917} | xe-4/0/7 | asw2-a5(ex4500):xe-0/0/4 | Yes | Will utilize current fiber -- downtime may be up to 1min |
| cp1075 | xe-4/0/11 | asw-a:xe-4/1/2 | Yes | Will utilize current cable |
| cp1076 | xe-4/0/13 | asw2-a5(ex4500):xe-0/0/5 | Yes | 5M DAC #4888 |
| dbproxy1012 | ge-5/0/7 | asw-a:ge-5/0/7 | No | |
| labstore1008 | ge-5/0/14 | asw-a:ge-5/0/14 | No | |
| db1116 | ge-6/0/11 | asw-a:ge-6/0/11 | No | |
| db1066 | ge-6/0/16 | asw-a:ge-6/0/16 | Yes | Downtime needs to be as minimal as possible |
| labstore1009 | ge-6/0/17 | asw-a:ge-6/0/17 | No | |
| dbproxy1013 | ge-6/0/32 | asw-a:ge-6/0/32 | No | |
| ms-be1040 | xe-7/0/29 | asw2-a5(ex4500):xe-0/0/6 | No | 5M DAC #4889 |
| cp1077 | xe-7/0/30 | asw2-a5(ex4500):xe-0/0/8 | Yes | 5M DAC #4990 **updated port number |
| cp1078 | xe-7/0/31 | asw2-a5(ex4500):xe-0/0/15 | Yes | 5M DAC #4991 |
| db1118 | ge-8/0/2 | asw-a:ge-8/0/2 | No | |
| torrelay1001 | ge-8/0/3 | asw-a:ge-8/0/3 | No | |