The Plan
My next step is to set up my e-mail server. I want that VM to be on the SAN. So I need to get iSCSI going. Also I need to get the home directories working, so it has a place for the .forward files and the inboxes and etc. I’m concerned that the new drives won’t match the drive that’s already in the SAN, so I want to do that first before I muck about with the configuration.- Install the drives in the SAN
- Configure iSCSI
- Get home directories using automount working
Installing the Drives in the SAN
I installed 2 more drives and put screws in the drive that was already in there plus the new drives, then reseated them.
Powered it back up and then had kinda a start, the power light started blinking rather than other stuff coming up. It did it for a long time, long enough that I thought huh, this is long, and then googling it. Long enough to find out that the ‘blinking blue light of death’ was a thing, and it meant a dead motherboard. Long enough to think Oh Shit, I can’t afford another one of these. Then it went beep and all the happy lights came on. Whew…
The new drives are indeed the same size so everything was easy peasy
Configure iSCSI
Oh hey, lookie here how to integrate a synology nas in your vmware lab that’s exactly what I’m doing.He’s suggesting using file based iSCSI. His reasoning is that the synology is used for other things, as is mine. However it sounds like the target is per volume and I believe I can configure this to be multi-volume. I’d like to try it, however it is busy expanding the storage pool and syncing the raid which sadly has greyed out everything. So, file based it is, we can revisit it later. And it looks like I don’t even get that choice since I got the cheap SAN.
I just followed his instructions and created the iSCSI target.
Now onto the esx server.
Looks like the esx server isn’t visible anymore. I believe you have to set up networking from the console anyway. Switched the cables around from the laptop to the server and that was messed up too. Grrr… Next thing I’m going to try is to put the dhcp and dns back. Or… actually before that find out what heartogold was set to before. Yeah that works 192.168.42.3.
I added the A record for heartogold to the new dns server. Then I got annoyed because it insists on putting up an error that you have to acknowledge every time you update the screen. Several of my devices have spaces in the name they’re asking for from dhcp which fucks up dns.
Anyway, the TVs had spaces in the name, I went to my.roku.com and changed the name of the devices. The other thing a levitron switch I couldn’t fix, so I gave it a fixed address in dhcpd.conf.
And now for something really brain dead. I deleted the network the management console was on. Mainly because I named it deepthot.int, and I wanted to get rid of it. Foolish girl. So I ended up having to reset the esx server to its default and recovering the VMs. Turns out not to have been as painful as I feared, so yay. This time I think I’ve got it set up so that you can change the IP from the console. That was greyed out which didn’t help matters any. Anyway what I did was:
- From the console reset the configuration
- After it booted set up the IP 192.168.42.3
- While I was going crazy the desktop machines lost their IPs and stuff since the vsphere was doing the dhcp. I manually set up the ip and set the dns to google.
- Went into the esx web page and recovered the VM and booted them up.
Then, and this took me awhile to figure out, you have to partition the iSCSI device. Go into the devices tab right click on the iSCSI disk that is there now, and select edit partition table.
Then it shows up as a datastore. Huzzah
I didn’t grok that you needed to create a partition table. I just set up the target and wondered why it wasn’t showing up.
I think I’ve done enough for today, especially given that disaster. So tomorrow automount I think.
Haven’t gotten very far, way far from automount at this point, there’s a whole bunch of freeipa stuff to set up before I get to there. I’ve forgotten most of it. Anyway…
To set the hostname with dhcp on bionic beaver you need to do more than install. You also need to edit /etc/cloud/cloud.cfg and change
preserve_hostname: false then reboot. And, of course on ipa you need to killall -HUP named-pkcs11 and on the laptop sudo systemd-resolve --flush-caches

