@ScottPack Depends on the size of the university. In smaller ones, that's probably the case, in larger ones I think staff are treated more like corporate employees. It really is a crapshoot though.
@MarkM I think the phrase, "treated like corporate employees" is the key phrase there. That has a lot more to do with management ideology than size, though.
@pauska That doesn't really achieve the same goal as tiering.. intent log and read cache are one thing, but it doesn't let you guarantee that oft-accessed data is always in flash
@TessellatingHeckler They were never very open with what features would be implemented next. They did discuss ideas, and I think community feedback drove priorities a bit. But the community rarely knew more than a feature or two ahead of time.
@Holocryptic At lunch I read a scene where one of the characters is walking by him while he's asleep and is made uncomfortable because he wakes up, looks at her, then closes his eyes without anything else changing. Fun stuff.
@ShaneMadden Right; you can't add a bunch of 3TB 7200 drives that'll be used for snapshots and other uncommonly used data; then add a bunch of 15k SAS drives which will automatically be used for "current" data. The ZIL and L2Arc are useful, but in the context of a whole pool only.
@Holocryptic Earlier this year I started at the beginning, with the notion that the last book would get published before I'm done. It has reminded me exactly how mediocre an author he is, and how much I wish Sanderson took over several books earlier.
@ScottPack I agree with you to an extent. But I can't help but feel like the ending is rushed. There was all this tension and buildup, and now it's going breakneck to the finish. Maybe I'm just sad it's ending...
@Holocryptic And by "tension and build-up" you mean 5+ books where nothing happens except for drawing out the story as long as possible to create more books :)
@Holocryptic It's kind of like the 5th season of B5. They had all these plot threads they wanted to finish out, but none of them were connected to the main story arc (completed in S4), and not really a seasons worth of them to begin with....
@voretaq7 Well, that and the impending cancellation at the end of S4 forcing them to wrap up the main story arc then, as opposed to letting it play out more slowly through S5.
@voretaq7, what is wrong with saying they use SQL. Everyone sticks to pure ANSI SQL that is portable to any SQL engine rright? All those vendor extensions don't add anything useful.
Totally not related to anything we're talking about but I just had an amazing thought: Make ice cubes out of coffee, so that when they melt in your iced coffee, it doesn't water it down, it just slowly replenishes it.
@MarkM You have two options. 1) Keep the answer up and then hope that I get more than 10 upvotes above your answer so I get the gold badge. 2) Spite the OP and delete the answer because he's clueless. =P
Monitor the event logs on the domain controllers for event 672. It will show all successful logon events for AD users and which computer they logged on from. Filter by computer to get logon events for only those computers that you are interested in.
@pauska Not if ZFS could tell the HBA how to calculate parity/crc/etc... Just saying, it should be entirely possible to create a HBA API where a volume manager can program how the HBA operates in regards to RAID.
The only time I delete anything is if my answer is potentially dangerous and I didn't realize it at the time of writing, which has only happened once from what I can recall. Or if I totally misunderstand the question and spout off nonsense :)
@voretaq7 I have a virtual environment with ~5 vm's. I would like to make another vm that only runs ssh (on a non-standard port). I will disable root login, and only have one user with a ssh account. I'd like to use this setup so I can ssh in from a remote location. I'm assuming that by removing shell access for this user, things are more secure
Not sure that disabling a shell will really make a huge difference. The bigger point would be to disable password authentication, and make sure your system is locked down.
but why are you setting up said tunnel? You can do it, sure - just do like Zoredache said (you can even do it in your authorized_keys file for a specific key)
"I am receiving complaints from users about wanting to see that accessing any links mysite.org/xxx/xxx be automatically redirected to mysite.org/xxx/xxx" -- I read that too fast and thought I had a client who did porn hosting with the new .xxx TLD
I use SSH tunnels to access our monitoring server. I do it because I SSH into that box regularly, and it's easy to just set up -L8181:127.0.0.1:8181 and not have to VPN :)
then here you are asking "I have a protocol for SHELL access, which I want to use as a TCP ROUTER, can you tell me which parts of which config files to bodge to make that sort of work well enough"
@TessellatingHeckler that's unfair to SSH -- the port forwarding is pure hackery, but SFTP/SCP are actually defined protocols that can be activated independently of shell logins
it just happens that we bodged them into sshd because... well idunno, ask Theo. He's the nutcase OCD security badger.
hmm, it's 20:11, wasn't I supposed to have something to do at 20:00hrs? ... ah yes, there's my "reminder" from Outlook hiding behind all my other windows.
Anyone have any advice for troubleshooting "The naming context is in the process of being removed or is not replicated from the spceified server" when attempting to force replication?
@ITHedgeHog 's the only time I ever saw that error - we had a "stale" DC entry that hadn't existed for like 2 years but never got demoted/removed from the domain properly
pfSense has FreeBSD at it's core. FreeBSD has physical interfaces (real ones connected to a NIC and fake ones for VMs), you then connect those to a vLan interface, or link aggregation (teaming), or bridging, or whatever sort of topology you want to create.
@voretaq7, it completely depends on the setup. Assuming your domain has multiple DCs, and all the others are working, and the DC with problems is not providing any other service, then a demote/cleanup/promote isn't that bad.
@SpacemanSpiff if you don't see the ARP (or its inconsistently getting dropped) that's a starting point - the fact that it's inconsistent is strange though
@ITHedgeHog oh no they did worse than that. The machine died so they just wheeled its corpse out the door. AD thought it was still a domain member too.
The funny part of AD is that there are times your replication can get screwed up, yet it'll seem to keep working. You just get gremlins on the network.
AD is like a little child: It sees various things on the network, grabs them and shouts "MINE!", then pitches a fit when anyone else tries to touch them.
@hobodave IP addresses don't belong to physical interfaces, they belong to some virtual idea of a network connection point. You can have two interfaces in a bridge with one IP, or in a team, or one interface with many VLANs and no IP
now, those broadcom nics in the other vSwitch go to a different blade fabric switch, but the configuration of the two fabrics is identical, and go back to the same juniper core
@TessellatingHeckler, oh I see, UK power. I never have understood why there is no standardization of power connectors across the world. Didn't recognize it, i am used to a NEMA 5-15.
Is that from the US? Perhaps it is just that the picture isn't good, but the non-rj45 end looks like a cylinder. In the US that might go to the ground pin, which shouldn't kill anything assuming the building is wired correctly.