From: Michael T. <mic...@ho...> - 2018-02-15 23:24:04
|
It's me again with a different question. Through the years one of my chunkservers have gone through several disk replacements, upgrades, and additions. Disks are mounted in /mnt and named mfshdd<seq> with <seq> being a through z, ie. mfshdda, mfshddb, and so on. It started out with 3 disks, mfshdd[a-c], and now it is up to the letter i. mfshdd.cfg now looks like: /mnt/mfshdda #/mnt/mfshddb /mnt/mfshddc /mnt/mfshddd #/mnt/mfshdde /mnt/mfshddf #mnt/mfshddg #mnt/mfshddh /mnt/mfshddi I'm toying with the idea of changing a couple of the mountpoints so that I have no holes in the sequence. The steps I'm thinking of is: (a) stop moosefs-chunkserver (b) remount mfshddi as mfshddb, and mfshddf as mfshdde. (c) edit mfshdd.cfg to reflect this change (e) start moosefs-chunkserver Is this doable without losing out the chunks that are expected to be in mfshddf and mfshddi? They're still there anyway, just in a different mountpoint. Will they be recognized as valid chunks when the chunkserver is started back up and it scans the folders? --- mike t. |