The following hosts have been replaced in s4, which hosts with larger disks:
eqiad:
[] db1081
[x] db1084 (moved to s1)
[x] db1091 (moved to s1)
[x] db1097 (moved to m1) T254556
[x] db1102 (currently backup source > moved to x1)
[x] db1103 (moved to x1, to replace db1127 so db1127 can go to s7)
[] db1121
[] db1138
codfw:
[] db2090
[] db2073
[x] db2091 (moved to s8)
[] db2099
[x] db2084 (moved to s8)
They should be relocated to other places including:
At least:
- extra hosts in s1 DONE: db1091
- extra hosts in s8 (including extra vslow): db2084, db2091
- extra host in s7 DONE: db1127
- Maybe replace x1 hosts with these ones (smaller disks) DONE: db1103
- One host must go for backup testing to replace the one that was taken: most likely m5 master after all the upgrades
Other movements
[x] db1127 from x1 to s7
[x] db1135 from m1 to s1 and db1080 to m2 (finally db1080 will be moved to m1 because of T256717)
[] db1132 to m3
[] db1128 to m5
[] db1133 to backup testing
[] db1107 from s1 to m2