11.3 release - offline and replace process now slightly different?

[Greg_E]

Note that I did not check for new procedures in the documentation, I just assumed it was the same as previous versions. I just loaded up an 11.3 for my XCP-NG lab and found what may or may not be an issue. This is an old HP DL360e G8 server with the onboard SAS RAID controller in AHCI mode (to export disks as JBOD), 12 cores and 24GB of RAM. Started with 4 old 250gb drives because it was what I had, and decided this wouldn't be enough. Spent last night reorganizing my portable drives so I could pull out 4 old 500GB drives so that I could make the pool larger. The process of going to pool --> status and offline --> replace drives is something I've done a few times on other machines, but this one threw me for a loop.

I offlined the first drive, stuck the larger drive in, and tried to replace. The "fresh" drive didn't show up as available. Tried the handy "refresh" button in the top right corner of the web interface which then replaced the previous drive ad0p1 with a string of numbers. Tried the replace function again and still nothing. right clicked on the disk menu to open in another tab, checked and the disk was visible and marked an unused. Went back to the pool status and scratched my head (and pressed the refresh button several times). Ended up clicking to a different menu and going back to pool --> status and then tried replace again, and now the replacement drive appears and the process proceeds. The next three drives were done by offline, remove drive, add drive, go to disk menu and wipe drive, then go back to pool --> status and replace drive. I'm fairly sure I didn't need to jump around like that in 11.2-u7 which was the last time I upgraded a pool to larger drives, so this seemed odd to me and I thought I should mention it. Not sure if it is a drive controller, drive, or software issue. There were no issue resilvering the new drives, everything worked as expected once I figured out that I needed to go to another menu before "replace" the drive.

I did have another issue because I was trying to extend a second pool, but I'm not sure what is happening there. The second pool is on an old 9650SE controller and it's not happy in JBOD and seems like it won't hot swap. I had to reboot for it to see the new drive I was going to add. I also can't get that pool to extend from 3 drives to 4 drives, but I can add the fourth as a spare. Going to let this sit until I get my new (to me) p420 controller which should put me all on the "HP" ecosystem and will hopefully support the chassis better and give me back JBOD hot swap.

 

[Jerami1981]

I just happened to need to replace a drive as well. I used the guide for 11.3 Release as i hadn't replaced a drive in over a year. I took the bad drive and set it to offline. Went and replaced the drive. Went back to my computer and tried Replace, but no options came about. I tried refresh a few times with no results. I thought i might have had a bad disk, so i went and put a new disk in, and got the same results. However, something in the system pooped out during this process and the server rebooted. When it rebooted, the new drive showed up to replace. At that point, I shut it down, put the first replacement disk back in, powered it back on, and was able to complete the process with the drive I originally intended to put in. I don't know if the reboot was required, or if i needed to wait 5-10 minutes for something to scan the disk before it was available, but the process definitely seemed different than the last time i did it.

 

[Greg_E]

Thanks for the reply. A reboot definitely wouldn't make sense on a system like this. I have a feeling that the interface just doesn't rescan the drives like it did in previous version. I forgot to check the terminal to see if it saw the drive change, but I'm guessing the system really sees the drive, just that the interface doesn't refresh to load the new options. A pretty minor issue but one worth noting.

I'm going back to 11.2 for a while, I just can not get SMB shares to allow me to get into them. I even set the everyone user to have full control and still nothing. Some times I can get a log in box. But this is for another thread and my time to play with it and figure out what I'm doing wrong is pretty much gone, need to get on to XCP-NG and figuring a bunch of stuff out with it. I may try 11.3 after it gets a couple of updates, that's my normal routine and I only broke from it because I was setting up this testing machine and thought it couldn't hurt.

And bringing it back around to the original topic, I had a couple of drives to change out again, I pushed a couple back into service that had some notes about failures written on to them. I'll try swapping them under 11.2 and see if thing go the way I think they should.

 

[Jerami1981]

Knock on wood, but hopefully i don't need to replace another drive for a while. When I do, I will try and remember to report back here on the experience.