|
adorai posted:it's ELK stack. Elastic is just one component. Not according to the people that maintain it.
|
# ? Aug 23, 2018 03:40 |
|
|
# ? May 9, 2024 09:03 |
|
Actually, they renamed it. Now it's called Elastic Stack.
|
# ? Aug 23, 2018 03:40 |
|
adorai posted:it's ELK stack. Elastic is just one component. It's Elastic stack, Elasticsearch is just one component but they've expanded beyond the core ELK offerings. I know I sound like a #brand representative but beats was the big thing that prompted them to drop the ELK acronym, and beats is pretty cool. Especially if you need to aggregate container logs. Our Elastic infra is nuts, due to a dumb error with our connectivity between one of our logstash clusters and its Elasticsearch cluster we weren't pushing logs for almost an hour a few weeks ago and we ended up with nearly a TB of queued logs in our logstash cluster. As soon as ES came back we had to scale a bunch of extra logstash nodes to handle the backlog. I'm just starting to get comfortable with some of this stuff and am constantly blown away by some of the scale we're doing at my newish job.
|
# ? Aug 23, 2018 03:50 |
|
Since you mentioned containers, if anyone happens to have a primer that's better than "run these commands and then presto you ran a container" I would be grateful. I get the overall idea and concepts but stuff like "just run docker-compose up" isn't very helpful.
|
# ? Aug 23, 2018 03:56 |
|
It kind of depends on what you're doing with them. Since I generally administer container fleets that somebody else has built, my interest is mostly in orchestration via kubernetes or ECS in which case knowing the commands to bring up groups of containers and how to get logs out of them via either Beats or just from the cli is all you really need. 'Up and Running with Kubernetes' is a good book for learning kubernetes basics. Katacoda has some good labs too, there's a bunch of stuff out there. There are repositories of most common floss products available so you usually don't need to build bespoke containers unless it's for software you're developing yourself. I'm no expert but I leaned by trying to containerize a basic web app a friend had built. It involved getting an nginx container to talk to a php container, nodejs container, and a mariadb container. I'd never deploy something like that to production - database containers are questionable in my mind (although I know some people swear by them). But it was good practice. I just worked from the official Docker docs and by pulling apart other people's Dockerfiles to see how they were building different layers. 'Docker Deep Dive' is a good resource if you really want to understand Docker.
|
# ? Aug 23, 2018 16:08 |
|
Database containers are fine, but anyone running them needs to understand the runtime requirements and maybe not run with as much dynamism as you might run a stateless application instance, because there's a lot more at stake if something goes wrong. Automatic failover and self-healing are great, but a lot of us olds are still wearing the war wounds of a MySQL instance that flapped too fast in a cluster capacity and trashed the whole DB volume.
|
# ? Aug 23, 2018 16:39 |
|
I'm lazy man, just RDS it and forget it
|
# ? Aug 23, 2018 17:52 |
|
BangersInMyKnickers posted:This is not true, it will opportunistically write logs to SD if it is large enough (I think 8gb+ is the delineation, but its been a while). Much appreciated— I’ll keep all this in mind when I figure out what to purchase. Option 1 sounds pretty promising.
|
# ? Aug 24, 2018 04:16 |
|
H2SO4 posted:Not according to the people that maintain it. Huh, well I guess I am just an old man behind the times.
|
# ? Aug 24, 2018 05:04 |
|
What’s the consensus on 6.7? Any gotchas coming from 5.5?
|
# ? Aug 26, 2018 17:41 |
|
Wait for 6.7U1 at least, unless there’s a feature you can’t live without and are willing to deal with potential instability. There are bugs and interoperability issues still that need to get smoothed out. You could pilot 6.7 in a test environment maybe. I wouldn’t run with anything newer than 6.5U2 in production for now.
|
# ? Aug 26, 2018 17:51 |
|
I've been running 6.7 on 6 hosts for a little while now, no issues.
|
# ? Aug 29, 2018 02:13 |
|
6.7u1 is coming later this year and has a feature complete HTML5 client so I would hold out for that unless you have a pressing need for another 6.7 feature
|
# ? Aug 29, 2018 04:11 |
|
So, wait for 6.7u2 for feature-complete, bug-free html5
|
# ? Aug 29, 2018 11:12 |
|
Number19 posted:6.7u1 is coming later this year and has a feature complete HTML5 client so I would hold out for that unless you have a pressing need for another 6.7 feature Maybe this will get my friend to shut up about the fat client. Maybe. Unlikely
|
# ? Aug 29, 2018 12:34 |
|
Small task/question: usually can be done in html5 or start typing in <perf and syslog tool here> More complex stuff not on html5 yet: Honestly I'd rather use powercli / orch Desktop client, post-2014: lolwut Potato Salad fucked around with this message at 21:49 on Aug 29, 2018 |
# ? Aug 29, 2018 21:46 |
|
We're running a 6.5 U2 across two sites with external PSCs. Now that 6.7 U1 has been announced, it's time to start thinking of upgrading. I'm pretty sure I want to do the equivalent of a greenfield deployment and stand up new 6.7 U1 VCSAs and then migrate my hosts from the old VCSAs to the new ones. Granted it's a little early, but do any of you experienced folk think that's a poor decision? It seems to me like it would add just a bit extra redundancy/safety during the move and I don't see any downsides...
|
# ? Sep 4, 2018 16:31 |
|
It's your decision to make, but if you're not just talking about VCSA you're going to want to validate all of your server hardware.
|
# ? Sep 4, 2018 16:50 |
|
Anyone gotten around to the L1TF/foreshadow patching? I'm getting things lined up, and not sure whether to start with the VMware infra, or the guest VMs. Both are internal, trusted, and not shared. We probably will decide to not kill HT at the ESX level, but that'll happen through a risk-acceptance committee - and I still have to get everything lined up anyway. I've got a couple of dumb questions that I can't comfortably answer: If all VMware patches are applied, do I really need to patch all VMs? If all VMs are patched, do I really need to patch vCenter and ESX? To make things extra fun, this is a UCS environment, so we'll pile on a couple of Cisco patches with microcode and the like
|
# ? Sep 4, 2018 19:14 |
|
I did my production ESXi hosts and VCSAs last week. It was pretty painless. The VMware KB seems to indicate that hypervisor level patching is independent of OS level patching, but I don't have any technical chops to really evaluate it.
|
# ? Sep 4, 2018 21:29 |
|
Yeah... that's what I'm trying to work out myself. Did you enable the new scheduler? If yes, what was the hit to performance?
|
# ? Sep 4, 2018 23:19 |
|
Happiness Commando posted:We're running a 6.5 U2 across two sites with external PSCs. Now that 6.7 U1 has been announced, it's time to start thinking of upgrading. Why does this feel safer? The upgrade involves you deploying a new VCSA and then running the upgrade tool to pull in the DB information. Your old VCenter doesn’t get deleted or anything. If you need to roll back you just power it back on and power off the new one. Take snapshots of everything and just do the upgrade. It’s pretty painless.
|
# ? Sep 5, 2018 05:05 |
|
Alfajor posted:Yeah... that's what I'm trying to work out myself. We are so comically under utilizing CPU it's not really relevant at all to us
|
# ? Sep 5, 2018 18:05 |
|
Happiness Commando posted:We're running a 6.5 U2 across two sites with external PSCs. Now that 6.7 U1 has been announced, it's time to start thinking of upgrading. Two things that you need to keep in mind - 6.5 and 6.7 have different host support (vcsa 6.7 cuts away esxi 5.5 support) - unless your db data is hosed up or you don't have any idea if there is any custom settings modified by someone else, there is little gain to do a full wipe rather than upgrading the vcsa payload using ami. If you are afraid of data loss, just do a full snaphot-based backup of each site vcsa before upgrading(you need to upgrade psc before the vcenter services node) SlowBloke fucked around with this message at 22:25 on Sep 6, 2018 |
# ? Sep 6, 2018 22:22 |
|
Happiness Commando posted:We are so comically under utilizing CPU it's not really relevant at all to us We're not too bad, but there are bursts that make compute go over 60%. That, when taking into account DR scenarios where all workloads run in 1 of our 2 datacenters, has been the greatest pain in trying to figure out what to do. But, since we run workloads that expose services to the web, and we're under lots of regulatory compliance needs, looks like we'll end up patching all guests, and VMware (but probably not enabling the new scheduler). some good times ahead.
|
# ? Sep 7, 2018 18:58 |
|
It's TYOOL 2018. Why is making an ESXi bootable USB any more difficult than dd if=hypervisor.iso of=/dev/rdisk2 on my Mac
|
# ? Sep 7, 2018 21:10 |
|
Martytoof posted:It's TYOOL 2018. Why is making an ESXi bootable USB any more difficult than dd if=hypervisor.iso of=/dev/rdisk2 on my Mac
|
# ? Sep 7, 2018 21:59 |
|
I've always just installed it from the virtual media support on an iDRAC/iLO
|
# ? Sep 7, 2018 22:00 |
|
Thanks Ants posted:I've always just installed it from the virtual media support on an iDRAC/iLO ^ Most orch tools or their communities have boilerplate for automating ipmi media, even
|
# ? Sep 7, 2018 22:59 |
|
SlowBloke posted:If you are afraid of data loss, just do a full snaphot-based backup of each site vcsa before upgrading(you need to upgrade psc before the vcenter services node) Do you know if there's a good and easy way to transition from external back to embedded PSCs as part of the upgrade? They seem like an extra added infra annoyance that I don't want to keep around if I don't have to
|
# ? Sep 8, 2018 18:15 |
|
Number19 posted:6.7u1 is coming later this year and has a feature complete HTML5 client so I would hold out for that unless you have a pressing need for another 6.7 feature I have the gear for RoCE support on a new cluster, which was my primary interest in 6.7, but we can wait until U1.
|
# ? Sep 8, 2018 22:10 |
|
I’m in the process of planning out a 6.0 vCenter windows server to VCSA migration. What’s the best current release of vCenter 6.5 to move to? I’ve read that the newer releases don’t currently have migration paths to 6.7, which is a bit disconcerting.
|
# ? Sep 8, 2018 23:03 |
|
Spring Heeled Jack posted:I’m in the process of planning out a 6.0 vCenter windows server to VCSA migration. What’s the best current release of vCenter 6.5 to move to? I’ve read that the newer releases don’t currently have migration paths to 6.7, which is a bit disconcerting. https://vspherecentral.vmware.com/t...ter-server-6-7/
|
# ? Sep 8, 2018 23:05 |
|
Happiness Commando posted:Do you know if there's a good and easy way to transition from external back to embedded PSCs as part of the upgrade? They seem like an extra added infra annoyance that I don't want to keep around if I don't have to The migration wizard or a VAMI update would keep them separate anyway. If you want to move from a External psc to a Embedded one I'm afraid starting from scratch would be required, all VMware docs i've found explain how to move to Embedded to External but not viceversa
|
# ? Sep 9, 2018 11:22 |
|
SlowBloke posted:The migration wizard or a VAMI update would keep them separate anyway. If you want to move from a External psc to a Embedded one I'm afraid starting from scratch would be required, all VMware docs i've found explain how to move to Embedded to External but not viceversa VCenter 6.7u1 has a tool to migrate external to embedded called the VCenter Server Convergence Tool. https://www.virtualizationhowto.com/2018/08/vmware-vsphere-platinum-and-vsphere-6-7-update-1-released-new-features/
|
# ? Sep 10, 2018 02:21 |
|
YOLOsubmarine posted:VCenter 6.7u1 has a tool to migrate external to embedded called the VCenter Server Convergence Tool. Good to know, VMware docs are stuck at 6.7 RTM so there is no mention of that tool there.
|
# ? Sep 10, 2018 07:00 |
|
Is Disk2vhd a practical free way to keep a copy of a (physical) Windows 7 machine that I'm scrapping but which I'll occasionally want to spin up on a new Windows 10 machine? I see Macrium and Acronis have solutions as well, and I'm not sure what there is to recommend one option over the others.
|
# ? Sep 10, 2018 14:47 |
I ran into an issue where a CentOS 7 VirtualBox VM on a Windows 10 host kept timing out trying to access mirrors.fedoraproject.org. I was stumped as to why the same VM worked fine on my machine (same version of VirtualBox, Vagrant, Windows patch level, etc). A curl of mirrors.fedoraproject.org from the host machine worked fine, and times out on the VM. I ended up coming across this stackoverflow post and the natdnshostresolver1 solution fixed the issue. I am completely stumped as to why this was only an issue on someone else's machine though, and why it works fine on my machine. Some Windows Firewall setting I've missed? Anti-virus mucking things up?
|
|
# ? Sep 14, 2018 22:11 |
|
Why does VMware make it so difficult to unmount a datastore & detach a LUN across multiple hosts? Unmounting a datastore they've got down as bulk unmounting is fairly easy, but detaching the LUN is still a tedious task of going to each ESX host, finding the LUN and manually detaching it. I've only found 1 or 2 scripts that can do the same thing, but they all jump right in and try to do things like unmount all datastores on a host, or they only do it for clusters of ESX hosts. Guess I have to write my own, or something *sigh*
|
# ? Sep 15, 2018 00:27 |
|
|
# ? May 9, 2024 09:03 |
|
Wicaeed posted:Why does VMware make it so difficult to unmount a datastore & detach a LUN across multiple hosts?
|
# ? Sep 15, 2018 03:54 |