Log In
Register
Home
Topic RSS
Read the original blog post
12:46 am
August 18, 2011
Offline
LWK texted me a reminder today of a request a while back whether we could put up a Vanilla world. I tried finding the original post just now with no luck; then again, I am rushing to post this. It was probably right in front of my face.
Mudwog and I discussed putting up a bare Spigot server at the beginning of the year; something that would allow members to play with close to home. We have had bare servers before – one pre-Arda and the pseudo-test server last summer for several initial 1.9+ updates. What we recall was that members initially explored and played around with new features, but then usage fell. The most common comment was that members missed our mainstay plugins and that having come and seen, they would be back when updates were done.
We considered making it more enticing this time around, creating a separate server with the new Seraph’s Realm seed, deployed with as many plugins as possible. We thought to merge this world and its content into the final update. Ergo, players would not lose their builds, inventory, currency, or other “stuff” provided by plugins. However, as nice as it would have been, the amount of work required to merge claims, balances, shops, etc., would be impractical. We would also need a master whitelist for preventing John/Jane Q. Public from joining and mucking things up in one way or another. Furthermore, we felt that some members may feel chided not knowing or having the opportunity to “pre-explore”. Hence, we didn’t consider it any further.
However, for some reason LWK’s reminder was still floating around in my mind all evening long, spawning other points of interest and concern wrestling with each other, till I came to the following. Before we can update Arda, we need a test server for live trials of all plugins, new and old. While there would be setup work, there is no show stopper that prevents us from putting up a pre-release of the test server now!
Preliminary Test Server Talking Points
Let me present the following points and solicit your comments/ideas. Please keep them simple and avoid splitting hairs. This is after all intended as a test server, a.k.a. temporary server, and not to be set in stone. The primary goal is testing all plugins while at the same time making it as reasonably possible enjoyable to all.
1. Do we use the new Seraph’s Realm seed?
Personally, I am for it, as it would allow members to explore and note interesting places for potential sites after the update. Should I be concerned that some members might feel slighted if they found out after the fact that they could have pre-explored the world?
2. Do we restrict access?
Given that a server whitelist is now UUID-based, we’d have to write a few scripts to parse our database to collect a candidate list of active members after some cutoff date, say 01/01/2016. Moderators can always add individual members manually should they have been missed; however, I’d like to minimize that effort and not have them swamped with requests. Any thoughts on how to handle this better? Or, do we try no access control at first?
3. Do we merge data into the update?
We will not merge non-world data into the update. Non-world data is any data tracked outside of a Vanilla world. This would include any plugin data and any player data such as player inventory, experience points, and other player-specific information. The only data merged would be the world itself along with items inside Vanilla containers such as chests, furnaces, droppers, etc. Members will be responsible for placing items in their personal inventory and/or shops into chests prior to the final update or lose them forever. Note: This only applies to the test server and does not affect the main server in any way. UDPATE: Furthermore, this only applies to the test server's Overworld; the Nether and End will not be included.
Once updates are completed, members can hoof it back to their Seraph’s Realm base (make sure to note its location) to find their items intact. Since all other data is gone, member bases will not be protected, leaving it up to them to move their stash elsewhere or create a new claim around it. Short of the honor system and fear of being banned, nothing would prevent a malicious player from stealing items; though, consequences of such would certainly ensue.
4. What about lost “stuff”?
This is a test server. There may be a chance that data and items may be lost due to crashes, resets, and anything else. The safest place to keep items are in chests. Anything else, including player inventory, is fair game to gremlins and such. While not guaranteed, claims are generally immune to loss.
Tentative Test Server
Without any feedback, considering it was just asked, these would be the operating conditions of the test server pre-release.
Initial Setup
- Uses Seraph’s Realm seed.
- Players receive into items when first joined (armor, tools, food, etc.)
- Players will receive 2,500 claims block and limited to 2 claims. Deleted claims are refunded at 75%. No additional claim blocks may be purchases. These settings allow members to claim one or two areas but also prevent them from going willy-nilly “reserving” spots all over Hell’s half acre.
General Operation
- Server will be up as much as possible.
- Server may shutdown or restart unannounced. We’ll try to give a heads-up, but not the usual 10-15 minute warnings. This is a test server.
- Server may crash and remain off for as long as required to determine/fix cause.
- A “Test Server” forum will be provided for placing announcements/change logs, bug reports, and feedback.
Bans are permanent and cannot be appealed; however, bans apply to the test server only and will not limit access to the main server now or after its update.- UPDATE: A ban applies to all servers, with no chance of appeal during the testing period. An appeal may be submitted after VC has completed its upgrade.
- Members are permitted to “test” plugins and Spigot for weakness, security, and/or potential exploits. These tests are limited to in-game attempts using a normal client with reasonable mods. However, members are bound by honor to report successes and not to exploit them (e.g. item duplication). Protocol and other server-targeted “tricks” are still not allowed.
- UPDATE: Unless specific sponsorship ranks are being tested, members will generally have access to Viceroy-level perks, even though their title/rank may be otherwise or none at all.
Updates
- There is no tentative order in which additional plugins will be added.
- Additional plugin command and usage documentation may be provided.
- Any plugin, except for GriefPrevention, may be removed without notice.
In Sum
I probably left out some important details while hurrying to get this “to the press.” Do not hesitate to ask reasonable questions. The goal is not to contemplate the best solution, but something that most would find reasonable and hopefully provide some willing victims, err… test subjects, to hammer out any kinks.
10:13 pm
Viceroy
Senior Mod
June 1, 2012
Offline
It's a great idea, frelling! I'd love to see it happen!
I think my biggest thought on it after thinking about it for a while is that consideration of non-Veterancraft players doesn't need to be too much of a concern. We certainly aren't advertising until the full update, so potential influx of new players would be negligible.
1) The SR being used for the map, that would then carry over to the updated server, sounds like the best plan. I would LOVE if we could use Arda, but that would require closing the main server so there wouldn't be two Ardas. Not really an option
2) A whitelist will be fine. Assuming I can edit it in-game, I have no qualms monitoring it. There may be a rush at first, but it would taper off quickly. And IF we can get some of our trusted players to be active enough again, some more staff would be needed anyways. We've talked about potential changes to the JM position that
3) If we whitelist, then I think there would be no problem with items being stored for eventual porting to the main server. We would just need sufficient time to let people know to start moving things into storage.
4) No worries there from me.
The setup sounds good, but if we're whitelisting then in general we should be able to trust everyone who will get on there, which means we should know them. I mislike the idea that they can do something to get banned on the test server but not the main server. For example, if Mr_Macintosh (sorry buddy) started griefing and flaming all over the test server, trust on the main server would be gone as well.
Anyways, good luck with this! Let us know if you need anything!
1:11 am
Viceroy
February 21, 2013
Offline
Oh yeah, you know me, when the rules go I'm making the server into my own apocalypse-themed party! In all seriousness though, I am really starting to like this idea. The test server that is currently up (or maybe down, I don't know at this point) hasn't been all too attractive to use, since anything we obtain there is not permanent. However, this idea allows us to collect new materials, explore a new world, and keep everything in the end, all while performing necessary plugin testing. It seems like a win-win for the server and for us, even if there are some hiccups along the way. If anything, running a server like this really couldn't hurt anything. If players don't want to use it, they don't have to, but now those of us who have been hungry for the taste of 1.11 can at least somewhat satisfy it.
I'd like to see this idea work.
Good luck with making that happen! And like LWK said, we're here to help if you need it.
If we don't study the mistakes of the future, we're doomed to repeat them for the first time.
4:40 am
August 18, 2011
Offline
@LWK
1) I had considered Arda, however, it is an older world and would not include 1.9+ items, biomes, etc. Furthermore, we will be running JVM monitoring tools which may be taxing at times with a 30K x 30K world.
2) Adding/removing player from the whitelist is very similar to banning using the /whitelist command.
4+) My thinking of the perma-ban on test server with no appeal was more focused on moderators not being distraction handling appeals. Obviously, we'd have to be careful about ban requests. For example, someone attempting to break into someone's claim is perfectly legit provided that they do it for testing purposes only. For example, if item frames are all of the sudden not protected they may be able to remove its item or even destroy the frame. They may try a couple more times, or call the attention of a mod, to see if it is repeatable. However, if they go nuts and break every single frame in sight and post a sign "I found a bug, rofl" that would not fly with me. If they can take items out of a container, no problem, as long as they put them back. However, you do have a valid point, if a player does something to receive the business end of The Ban Hammer, then they really did screw up. I agree that the ban will transfer to the main server.
4:57 am
August 18, 2011
Offline
Minor updates applied to the proposition: 1) revised consequences of ban; 2) added sponsorship perk info.
5:09 pm
Viceroy
Senior Mod
June 1, 2012
Offline
Fully concur with the update! I didn't even think about the biomes and items. Thanks!
5:33 pm
August 18, 2011
Offline
A couple of PM/emails ask pretty much the same thing
Not pushing, just curious. If you move forward with the pre-release, when will it happen?
Well, that all depends - doesn't it always. I mentioned that we are also looking at infrastructure in the addendum to this post. This weekend we are going to experiment with OpenShift, RedHat's value-added package of Kubernetes, a Docker image PaaS environment.
No, no, don't worry, we are not going to wrap Spigot into a Docker image. While the overhead is minimal, it doesn't have the same real-time flexibility as a normal server installation. However, other services that are mostly set and forget - when it comes to installation configuration - such as web servers, databases, mail system, wikis, etc. are perfect for Dockerization (if that is a term), allowing us to quickly add additional resources when needed.
Case in point, the test server needs a separate database system since we don't want to expose our production database system to potential test risks. Likewise, we may be using Redis or MongoDB for less database-ish type of data (e.g. configurations, preferences, logging, etc.). We are giving ourselves two man-days to investigate. If we find that we need to think about this approach in more detail, we will shelf it and create a regular deployment of services for the test server.
All in all, we hope to have the pre-release test server running by end of business Friday (5/5). It could be sooner if the results of experimentation are fruitful.
12:08 am
November 25, 2012
Offline
I like me some test server, and anything that helps speed testing along so you can eventually update. Anyway on to the points....
1. I get the reasoning behind making it seraphs realm and leaving it up post update of arda. I guess my concern is the wear and tear on the map happening so early. Current OR is pretty rough to find diamond ore etc on because its been there a long time, survived its share of xray people that were banned etc. I don't know how often in the future after the update we will be getting new secondary worlds for gathering stuff. Say the new seraphs realm has one or two big ice spike biomes, packed ices makes some great stuff and unfortunately doesn't seem they are going to be adding a recipe to craft it. Supplies of that type of block could be farmed or wiped out before we even have Arda updated as people could lay aside resources now for building on arda when it is up to date. Would it be possible to say, people go nuts on this version of SR, note interesting locales etc, stock pile stuff to move over, give us maybe a week window to transfer our stuff from chests once Arda is up and then give us the same SR seed/map but entirely fresh again?
2. Whitelist should be fine, perhaps a thread for people to post asking for permission who maybe haven't been on in awhile. I don't think people will be showing up and causing havoc who are brand new but I guess there is always a chance.
3. Being that it is a test realm I see no reason to merge data besides basics like items in chests, beyond that any xp gained/lost as it applies to basic xp or say mcmmo would be insubstantial or a worthy sacrifice to the test server in my book.
4. Again its a test server so its a enter, build, stockpile etc at your own peril, trying to replace things lost to bugs, crashes and wipes sounds like a lot of hassle and work for LWK and yourself, wouldn't bother with it myself.
Anyway, I look forward to a test server in the future to play on.
6:01 am
June 22, 2014
Offline
On a big enough map, wear and tear wouldn't be so bad. I do like the idea of staking claims, like on woodland mansions and the like, but again, that's just rewarding people who log onto the test server. Once the update happens, hopefully many people will return. Maybe those testers do deserve some kind of reward for their efforts. I've talked about my plans for the update in other threads, but I do reiterate the changes that will happen. Elytra flight will be here to stay, and there will be a limited number of elytra available to players, because the map size is limited. That Seraphs realm will likely have an updated End attached to it, meaning people can farm elytras there. The game changes completely when you have an elytra and a stack of fireworks. Running places suddenly becomes meaningless when you can fly dozens of blocks a second. And, fireworks aren't exactly difficult to make, but farming gunpowder is pretty tedious. I got about 15k stocked up for the update, with plans on making a store at New Glacia and selling people fireworks. Like I said, I would also like to get a healthy stock of enchanted elytra that I can donate to people coming back, because I don't feel that they should collectively be the exclusive reward to whoever gets the first one. After you get an elytra, farming them is the easiest thing in the world. You just get like 4 stacks of fireworks, an ender chest, and a silk touch pickaxe, find an end city, fly to the ship, grab the elytra and stock it in your ender chest and then pick it up. It's literally that easy. So, once they are gone from the End, they are gone forever. Unless, of course, Frelling is generous enough to reset the end for us when we run out, but I wouldn't place any bets on that. We haven't gotten End resets for other reasons in the past. The server selling them for shilling is also an option, and you can set the price artifically high so that trade among players is encouraged. I played on a server awhile back that traded in diamond blocks. The server price was 3x higher than the players were selling them to each other for though. I mean, that server sale option can exist if it needs to, but I feel players should in the end deal with other players for such an item. They will be crazy expensive at the start when the first ones are found, especially if someone no lifes for a few days and cleans out the End in that time. The other exclusive End item is the shulker box. These things are beyond useful, and the supply on those won't nearly be as limited as the elytra, but they are still a limited time item. I guess my main point, is the damage that just the End being open early will cause to the server. Someone making a base fully stocked with End stuff will clean up majorly when people come back.
This brings me to the discussion about plugins in the End. Traditionally this server has made the End plugin free, and perk free. You can't make homes, you can't teleport home, you can't do anything but play vanilla Minecraft there. I feel like we should be able to make claims on the outer islands, especially if we find a particularly beautiful End City. I have no idea what other priviledges that GP brings along with it, but that's something else that can be discussed. I guess that is all I have to discuss about the test world itself.
As far as whitelist is concerned, I don't have a problem with it. LWK knows most of the people that play regularly on VC now and in the past, and i'm sure he will pre-whitelist many people already. I just hope I'm on that pre-whitelist list, if it exists. =P
I believe my large paragraph above expresses my viewpoint on merging the data. I'm against it, ESPECIALLY if it merges including a pre-cleared out version of the End.
Bans should be dealt with pretty harshly in general, but I agree completely with bans including all servers VC operates. If a regular wants to log into the test server and grief other established players, the punishment for doing so better be pretty severe. Many players donate money to the server, and losing it all because they were stupid on the test serverwould be a bad way to go out. A justifiable way, but bad nontheless.
9:38 pm
August 18, 2011
Offline
WallyBean said
Current OR is pretty rough to find diamond ore etc on because its been there a long time, survived its share of xray people that were banned etc.
That is primarily due to the current OR having been the secondary world for the past 18 months or so. It is ancient in terms of secondary world lifespans.
... I don't know how often in the future after the update we will be getting new secondary worlds for gathering stuff
Secondary worlds generally are replaced with each major update. Thus, the next time will be whenever 1.12 is released, which may only be a couple months away.
Would it be possible to say, people go nuts on this version of SR, note interesting locales etc, stock pile stuff to move over, give us maybe a week window to transfer our stuff from chests once Arda is up and then give us the same SR seed/map but entirely fresh again?
That had been a consideration. While loosing builds wouldn't be much of a concern if this is known upfront, Mudwog did point out that a reset would be similar to a server-wide item spawn. Also with 1.12 around the corner, the new SR will be replaced with a new OR.
We also considered using a different seed for the test server now or for the SR later as part of the update. We are holding that as an option anyways, but would prefer to have some continuity.
10:10 pm
August 18, 2011
Offline
vyscaryska said
All good points, but mostly in regards of the main server, not the test server. However, I think I can sum it up.
Thank you for bringing up the The End, it made me realize that I forgot to add a clarification (see initial post for update). The Overworld, Nether, and End are three separate worlds. Only the test server's Overworld will be added to the main server. Any Nether or End Portals will wind up in the Nether and End of the main server. Arda's associated Nether (10Kx10K) will be wiped as the jump from 1.8.* to 1.11* includes changes thereto. The End will also be renewed, thus an opportunity to "find" more Elytras.
Given that there is more to do in The End, I agree that prior restrictions are too harsh. We are considering several options, many of which still depend on other tests and experiments. We will most likely make The End larger and take measures to make its longevity similar to the Nether; that is, it will only be wiped if an update requires a wipe. Core functions such as setting homes, shops, etc. will most likely be enabled.
However, none of those changes will happen at the update. We want to have more time to consider options and solicit opinions - not right now, but after the update.
We are also considering several approaches to replenishing hard to get items such as the Elytra. One option is to periodically spawn new ones that had been previously looted or in different chests all together. Another option is to add Elytra as a bonus drop with a very, very, very low drop rate. A third option is to add a custom recipe to create Elytra, although the ingredients themselves will take some effort to collect. There are other ideas, but as I said, it is too earlier to consider anything definitive.
2:16 pm
Knight
November 3, 2012
Offline
So many good thoughts. Fine, I'll throw my own on the stack too (in no particular order). And to preface the below quickly, please keep in mind I've beta tested quite a few games previously, so I have a pre-disposition toward trying to break things for the sake of testing, and I'm also good for playing devil's advocate... alot. With that in mind:
Bans: Absolutely should carry over to main world. We may not be able to carry over mcmmo stat increases and such from test/resource world to main world, but it looks like you're planning on making items transferable. And as item gathering also seems to be what gets abused the most (via x-ray, etc), attaching consequences to bad behavior that carry over into the main server is only logical and shows clearly that regardless of the world played on, we expect our community to be better than that and above that kind of behavior.
Elytra: Somewhat in agreement with vyscaryska, I think elytra are far too rare in vanilla considering the movement benefits and just plain fun-in-general they provide. I also agree with frelling that we don't want to make these items too common either because they'll absolutely lose their appeal quickly. I'm not sure if making them ultra rare bonus drops would work all that much as I think people would just start grinding whatever skill/ability whose drop table you add it to in order to farm them anyway - and farming leads to boredom very quickly. I think it makes better sense to make them craftable, but with ingredients that aren't easy to collect (as frelling suggested) - like what about a crafting table grid full of leather with a nether star in top middle slot? Those already cost a few thousand shillings on open market due to the effort it takes to collect them, and it would also make nether stars useful for something beyond just crafting beacons, too (giving additional benefit to another uncommon item already in the game). For most players who aren't good at killing withers or don't have l33t skills at end hunting, it would give them something to work toward / a reason to save up their shillings. I think that would strike a pretty good balance - making elytra uncommon but obtainable by the general population, while for those who do have the skills, providing a nice reward for putting those skills to use for the benefit of the general population. Win/win for everyone?
In regard to refreshing the resource world at any point after it's been created... well... I'll be honest, I think that would be a bad idea and would lead to all kinds of abuse issues. Herobrine whispers in my ear that I could run around the world and note coordinates for all of the rare/good stuff, and then when the resource world is refreshed, hit up all those collected coordinates and farm the good stuff all over again before others can find them naturally on exploration. Even worse, what if I note locations of diamond deposits, and then after refresh, dig straight to all of them to re-mine them? Technically, it couldn't be considered cheating (because I found them legit the first time), but still could easily be confused with x-ray to any admins/mods observing people digging straight paths to diamond deposits. With as much as you already have to deal with actual x-ray issues as is, I don't think you need that kind of confusion making your jobs as mods/admins even harder lol. Not to mention as frelling has already stated, resource worlds are just that, somewhat long term but still, in the end, ultimately temporary worlds that will be swapped out as versions of MC increase. So refreshing them isn't really needed.
Anyway, those are my thoughts to add to the pile. I look forward to jumping into whatever gets implemented! 
2:49 am
June 22, 2014
Offline
frelling said
I've been excited for this update for quite awhile. I definitely see myself coming back and playing quite a bit after the update. It's pretty important to me that everyone gets to enjoy it, which is why I want to do what my plan is with the elytra. Not having access to the end on the test world is more than adequate to prevent the update being spoiled for the end.
Making the End larger is definitely something possible. There are like 90% less blocks in the end than there are in the overworld, meaning you can make the world much larger with less server lag.
The test world looks like it'll be pretty fun. The last test world you had up was testing a bunch of new plugins like the NPC teleporters. I'm excited to see what else you want to add.
Trip6s6i6x said
The server typically doesn't regenerate the land unless it's been griefed or strip mined heavily. When Frelling talks about "regenerating" the resource worlds, he is talking about complete replacement. Every time the server updates patches, we typically get a new resource world, alternating between Seraph's Realm and The Outer Reaches. OR is pretty mined out now. I've resorted to going to the north west corner and just clear mining 3 blocks high on 11 for diamonds and gold. I'm stockpiling both for the update. We had like 2 stacks of diamond blocks, we now have like 6 because of that. It's annoying to keep moving the haste 2 beacon, but it really does speed up the process.
http://wiki.veterancraft.net/i.....ary_Worlds
As far as elytra, you mentioned them being rare, and that is kind of true. They would have remained so, because you can only glide so much with the elytra before having to land. it still made farming them a challenge. You'd have to pillar up and glide to new islands. or, make 1 wide bridges between them. The rockets being used as booster while flying changed that completely. Once you have an elytra, farming extras is beyond trivial. You don't even have to bother with raiding the city, you can just fly right to the ship and loot the elytra. And, once you have one, the only real way to lose it is to die. Feather falling IV boots make that pretty hard to do as well, unless you like fall all the way to ground from like 50 blocks up. You definitely won't see me with an elytra on unless I'm flying, or I have like 5 extra ones. An ender chest will become part of my usual roaming kit, and I'll keep an elytra in there. Everything has changed with the past few updates. The game is just a little more.... easy? idk.
1:19 am
November 25, 2012
Offline
I have some bad news for you Trip, diamonds are not worth much in current content. Herobrine is not your friend, he is just evil. As Vyscary hinted, the game is easier in its current state, for a number of reasons. Mending is probably the main one, you can find it in the End or trade a villager for a book but once you have mending on your armor/tools, you rarely need diamonds. Honestly the only thing you would accomplish by doing what you mentioned is wasting a large amount of time for virtually no reward, I mean unless you plan on building a large thing out of diamond blocks but I have yet to see anything that looked great that was diamond block heavy. Its essentially a decorative block on multiplayer servers these days. There honestly isn't a bunch of "good stuff" you are going to find to plunder again later. High odds you won't even find a forest mansion and elytra for people who like the oddball flying added to the game is your only possible valuable stuff and its not going to be on the SR map.
I will point out my suggestions are based on the huge raft of unknowns, so you can't fault me for not knowing if the server will actually be replacing resource worlds and updating frequently when the last 2 updates and resource worlds would suggest that is not going to be the case, at all. Be great if that is how it will be in the future but my points were simply based on past experience here, trying to help builders maximize what could be our resource world to use for the following year (plus).
6:19 pm
August 18, 2011
Offline
Everyone
While I appreciate the interest and enthusiasm debating finer points, this is not the place. To quote
Preliminary Test Server Talking Points
Let me present the following points and solicit your comments/ideas. Please keep them simple and avoid splitting hairs. This is after all intended as a test server, a.k.a. temporary server, and not to be set in stone. The primary goal is testing all plugins while at the same time making it as reasonably possible enjoyable to all.
All I am looking for in this thread are suggestions/concerns on the four Talking Points and the planned process for the Test Server as described above. I am certain everyone can understand that this is an important stage and anything that does not directly contributed to the test server and testing plugins would just be noise. If there are any questions or need for further description/explanation, I will ask.
That should not be construed that we don't want to hear from you. To the contrary, we encourage it. Once we are happy that the new infrastructure, server, and plugins work as expected, and a suitable rest period, there will be many more things to ponder.
7:54 pm
August 18, 2011
Offline
The "crash edumacation" on the finer points of Docker this weekend was a success. It has been at least two years since I last evaluated Docker and it has come a long way since then; to the point that I am updating infrastructure plans for work to leverage Docker for new non-mission-critical services to be deployed in a few months.
That said, we have added a little more work to our plate. This should not affect the pre-release test server plans, but may increase the overall deployment time of updates. We'll need to go slow initially, until enough anecdotal evidence has amassed to waylay any concerns.
6:44 pm
August 18, 2011
Offline
Things are going well. I like it when I'm learning new things or expanding on things I know. Both Mudwog and I have spun up on Docker changes since we last looked at it about 2 years ago. It has certainly grown in leaps and bounds from an infrastructure provider's perspective.
I have been attending webinars all day - more to come this evening - on OpenShift, a PaaS solution built upon Docker and other technologies. This is purely for research to note what is important to consider today, should we want to leverage OpenShift's capabilities tomorrow - those are metaphorical time references.
We had great success creating an OpenVZ 7 (Virtuozzo) and Docker hosting environment. More specifically, we are running OpenVZ and Docker side-by-side giving us both an OpenVZ container/KVM host and a Docker host. It appears that we are one of the first to do this. Everyone else seems to be focused on running Docker inside an OpenVZ container (i.e. VPS). We are extremely happy that this works as well as it does. We had anticipated having problems or spending excessive time tweaking things "just so" to make it work.
This will mean OS-level updates for our two servers and down time. However, since we want to try this first for the test server, we are looking into getting a temporary server provisioned by our data center and if they don't come through we'll deploy the system on AWS or Azure.
3:33 pm
August 18, 2011
Offline
We are in the process of deploying another server at our data center. This server is intended to 1) provide a platform for running the Spigot test server; and, 2) field test the design and configuration of our new PaaS/SaaS infrastructure. No need to get into the details right now. Server provisioning should be quick once we give the go ahead. Currently, I am waiting for a callback from OVH in regards to some technical details/clarifications specify to the system that we intend to deploy.
11:04 pm
August 18, 2011
Offline
Well... Server provisioning was a piece of cake. I barely had printed out the order receipt when I received an email letting me know that the server was ready for use; less than 5 minute turnaround time - nice! That is also where quickness stopped.
We've always used OpenVZ 6 to provision an OS container (a.k.a VPS) to host one or more Minecraft instances on our servers. Given that it was a package that installed on top of CentOS, deployment of new servers was always easy. Alas, Virtuozzo decided that OpenVZ 7 is now a distro. It is still CentOS 7, but has numerous tweaks and customizations added. OVH didn't have an image, forcing me to perform a manual install via IPMI. No biggie, we have VPS specifically for running net installs, rather than trying to upload a 1.2GB ISO. We had tested OpenVZ 7 extensively on our local servers, but never gave much thought to drive configurations, assuming it would be business as usual. So much for assumptions.
All our gaming servers have three 400+ GB SSDs, one for boot and host OS, and two configured as a RAID1 for OS containers- the mission critical data. We can always restore the host drive from an image. While we periodically back up everything offsite, RAID gives us extra protection against losing Minecraft servers and worlds in the event of a drive failure.
OpenVZ 7 will not allow us to use a RAID configuration. They claim that they have redundancy already built in and that a RAID, software or hardware, would just case delays when writing data. I don't give a ^#&*&%*#. These are high-end fast SSDs. If we needed more speed we could always opt for NVME storage, which takes data faster than any high-end server can write. Overhead due to RAIDing is in the noise. Occasional network lag will cause more delays than writing data.
Obviously, this is not so important for the test server. If it has problems, it will be because of plugins, not disks. However, given that we intend to update our other servers as well (during the final update), this is still of great concern. Fortunately, Mudwog was available and we spent the all of yesterday and into the better part of early morning today trying various workarounds to allow RAIDed drives. No luck.
As it stands right now, the only solution with OpenVZ is to performance incremental backups more frequently. We are still looking into more direct solutions such as Proxmox, LXC/LXD, etc. because this is really bugging us. I'm married to OpenVZ and love it dearly having used it for years on other projects; but it appears that there may be a divorce looming on the horizon.
We are going to spend the rest of today looking for better solutions. If that doesn't pan out, we will just live with it for the test server; though we will need to have a solution in hand before the final update.
Thanks for reading my rant. I feel better already. If anyone has any ideas for OS virtualization, not hardware virtualization (a.k.a. KVM, Xen, etc.), I'll gladly listen. Then again, we may just decide to run all our services, including Minecraft servers as Docker containers; problem solved 
6:23 pm
August 18, 2011
Offline
First things first - If you are a Windows user and you haven't updated, UPDATE NOW!
What is that they say about the "the best laid schemes o' mice an' men"? Go figure that someone decided to release a virulent virus - WannaCry; we had better things to do. Fortunately, the media did good getting the news out to the public; though lagging by almost 2 days after we got the first heads up that something was amiss.
The good news is that systems at work, our servers at OVH, and my systems are unaffected. 90% of them are Linux systems anyways and those few Windows boxes are always kept updated tout de suite. Unfortunately, I cannot say that about others. My phone has been ringing off the hook with panicked colleagues asking for help. That has been my life for the past 72 hours; running around with my arsenal of cyber tools in hand training and helping sysadmins restore locked systems.
The fires have died down mostly. I get no pleasure telling someone, "I told you so!" My favorite quote that adorns all my posts has rung truly loud these past few days. Some colleagues and perhaps a few ex-colleagues - given I was blunter than a 2x4 cutting stone - paid a heft tuition to the School of Hard Knocks. Backup, Update, Rinse and Repeat!
I'm closing loops and tying up loose ends, after which a short, but much needed downtime is required. Hopefully, I'll get back to the final configuration and deployment of the test server sometime later tomorrow.
Most Users Ever Online: 734
Currently Online:
24 Guest(s)
Currently Browsing this Page:
1 Guest(s)
Top Posters:
Emulated: 3206
ryanpitts: 1300
Dalferes: 747
Pherian: 660
Okarim: 594
Member Stats:
Guest Posters: 10
Members: 27638
Moderators: 3
Admins: 2
Forum Stats:
Groups: 8
Forums: 45
Topics: 6229
Posts: 27413
Newest Members: Qeegooni, JnerSnige, Lesliesak, ChongLog, Richuardkirty, Wujwcoorn, Albertrib, EwnDuede, DavidRep, Histdece, RonnieSpure, Nicholasto, AlbertTrona, Garry7Gap, Larrystork, Christianlh, bvidrZoowlz, GustavoWab, Crywoni, BillyNOR
Moderators: terrorisly: 424, mudwog: 127, LightWarriorK: 2154
Administrators: meatbawllz: 2475, frelling: 3264
