EVE Search - Server reboots are sooo....2004.

All Channels

Test Server Feedback

Server reboots are sooo....2004.

» Click here to find additional results for this topic using Google

Monitor this thread via RSS [?]

Pages: [1] 2 3 4 :: one page
Author	Thread Statistics \| Show CCP posts - 7 post(s)
Digicomm The Digital Communists	Posted - 2007.11.24 11:24:00 - [1] CCP, ever thought that maybe you have gotten into a behavior that this once needed reboot might no longer be needed now. I understand that back in the day your exhausted Hardware need a quick break. But there must be software driven utilities that essentially can do the exact same thing. An hour downtime a day is a bit much isn't it. Separate yourselves and do downtimes when needed. Sched reboot on Mon, Wed, Fri. or something...there has to be other options. FYI, posting pics of the CCP data centre would be nice too. :)
Eleana Tomelac Gallente Through the Looking Glass	Posted - 2007.11.24 13:25:00 - [2] It sounds like no downtimes would need much software developpement to do the same things while the servers are online. Many big updates are done in the database on downtimes (I can think of checking accounts, fixing messy market orders and returning the items, POS stuff, sovereignty checks, much cleaning of the space, exploration stuff, and many other we can't even imagine). Downtimes are not so bad, you can eat while it is downtime, sleep one hour, get social IRL... The fact is if it takes much time to develop, I prefer that we get new features instead of downtime removal... Or it would need a whole new team for a downtime removal project... It implies cost... -- Pocket drone carriers (tm) enthousiast ! Say hello to my tiny friends !

CCP Atropos	Posted - 2007.11.24 13:56:00 - [3] Edited by: CCP Atropos on 24/11/2007 13:57:09 The downtimes on the test server aren't just used to give the hamsters a break. It is a constantly evolving test server after all. As for TQ, it's one of the largest databases I know of, and the downtimes are used to do daily maintenance.

Kael D'mende Homo Victor	Posted - 2007.11.24 14:07:00 - [4] Originally by: CCP Atropos As for TQ, it's one of the largest databases I know of, and the downtimes are used to do daily maintenance. ... so basicaly just killing off old junk so the hamsters don't croke ;o) why this ain't possible to do on the "fly" is puzzeling me.. surely it must be duable to make it release old ID's etc during runtime... Regards. /Kael

CCP Gangleri	Posted - 2007.11.24 14:11:00 - [5] With a database as massive as TQ it does not make sense to risk faults or undue lag by trying to run large jobs while it is live. A 30-60 minutes downtime a day ensures more stability and that the jobs that need to be run don't interfere with anything else.

Corporati Capitalis Tollan Technologies	Posted - 2007.11.24 14:53:00 - [6] Originally by: Digicomm FYI, posting pics of the CCP data centre would be nice too. :) I doubt that'll happen to be honest. Collocation center operators don't really like it when people take pictures in their data floors and put them on the internet. And besides, CCP probably doesn't like it either.
Def Antares	Posted - 2007.11.24 14:59:00 - [7] Edited by: Def Antares on 24/11/2007 14:59:12 nm
ITTigerClawIK Amarr PIE Inc.	Posted - 2007.11.24 15:47:00 - [8] so when about is the test server comeing back up then? Sig space reclaimed in the name of me -courtesy of Tiggy ([email protected])
Chruker	Posted - 2007.11.24 18:33:00 - [9] Originally by: CCP Atropos The downtimes on the test server aren't just used to give the hamsters a break. It is a constantly evolving test server after all. As for TQ, it's one of the largest databases I know of, and the downtimes are used to do daily maintenance. I haven't seen Google doing 1 hour daily downtimes, and they must have a huge database ;-) ----- http://games.chruker.dk/eve_online ----- Top wishes: - No daily downtime - Faster training on sisi - Speedup IGB table rendering
Chi Quan DEFCON. Phoenix Allianz	Posted - 2007.11.24 18:44:00 - [10] there is nothing spectacular to a datacenter. a noisy, warm room full of racks and prolly blalecenters. -- Tempus fugit -- quote spiralJunkie: it doesn't matter how you pronounce it, it still shoots you in the face
BigWhale Gallente	Posted - 2007.11.24 19:25:00 - [11] Datacenter should be a noisy and COLD room. By no means a warm room. :) Well the warmer, the quieter... ;) More on the subject, the bigger the database, more often scheduled downtimes. You reall, really don't want indexes being built during uptime. I'd guess there are a lot of constaints, foreing keys and lots of tables that need to be indexed and rebuild each downtime. Can someone tell us how many tables are in the sql server? There are probably spread in more than one database? -- R, U & Y are letters...
Toa Amarr Orion Academy	Posted - 2007.11.24 19:34:00 - [12] Oh come on people the downtimes not that bad it's very helpful in some cases. I've rarly seen a 60min downtime for regular maintance and most of the expansions haven't been that bad after RMR. I was surprised at how soomthly the expansions for Rev have gone so far. I don't mind the downtime mainly because I'm American and the server is coming online when I get off work, but I can see how Europe players would have a problem as its in the middle of your lunch time. ---------------------------------------------- I dont get how this stuff works I just use it.
Adelorae24 Caldari Highwaymen	Posted - 2007.11.24 21:05:00 - [13] Far better to have the scheduled downtime than the unscheduled downtime that would result from trying to perform maint. on the fly. I guess the logistics might not be obvious to those who have never been responsible for the operation of a large scale database. At any rate, the daily downtime seems to be getting the job done just fine. If it ain't broke, don't fix it. New Sig Pending...
Helen Hunts Gallente Red Dragon Mining inc	Posted - 2007.11.24 22:13:00 - [14] Downtime is a good thing. I'd much rather have a nice, predictable downtime for regular maintenance, than irregular, edxstended downtimes for massive repairs/rebuild, or even worse, information loss and rollbacks. (although someone who just lost a Mothership would love a rollback) Besides, downtimes also give CCP a chance to peek under the hood to make sure that no shenanigans are going on with their servers. Plus, it's easier/faster to backup a server if it's not trying to run operations at the same time. CCP: Take your time, do it right. _______________________________ Mine da rocks, make more ships. Pop da rats, make more rigs. Sell da gear, make more money. Any Questions?
Cythrawl Caldari Central Defiance	Posted - 2007.11.24 22:22:00 - [15] Considering the amount of info in EVE, and the fact that they don't want anyone tampering with it from the outside who shouldn't be, the system is getting scanned every day for bugs and viruses and all those pretty little bits and bobs of into that can tell em there is a major problem. When's the last time you ran your virus checker? How long did it take? An hour a day to keep bad things away is pretty good if you ask me. Especially when the alternative is a computer game who's servers are f'ed. The fact they use solid-state memories makes their maint and upgrading easier than any other system. So CCP goes a good way out of the way to make sure the game doesn't fly off the end when we're supposed to have access, less the hamsters all decide its time to go into a 'highlander' moment and kill each other due to some misplaced integer. .:. Originally by: Krazy Bitsch Originally by: Virtuality In before the first troll. i do believe you are too late for this....
SoldierOfJustice Infortunatus Eventus HUZZAH FEDERATION	Posted - 2007.11.24 23:43:00 - [16] Originally by: CCP Atropos The downtimes on the test server aren't just used to give the hamsters a break. It is a constantly evolving test server after all. As for TQ, it's one of the largest databases I know of, and the downtimes are used to do daily maintenance. Originally by: CCP Gangleri With a database as massive as TQ it does not make sense to risk faults or undue lag by trying to run large jobs while it is live. A 30-60 minutes downtime a day ensures more stability and that the jobs that need to be run don't interfere with anything else. Rofl, those must be the newest jokes CCP came up with . Seriously, we've been playing long enough to know that all the so called maintenance you're doing during DT is just bull****. EvE server did have huge problems running the 23h a day its supposed to be able to do not long ago. And if am not mistaken most of the many unexpected downtimes were caused by database problems. So many other systems in the world have alot bigger databases than the one EvE uses, and they dont need dayly downtimes. The problem isn't the size of the database but the code your server is running. The code you devs are responsible for.
Eka Maladay	Posted - 2007.11.25 00:50:00 - [17] Originally by: SoldierOfJustice Rofl, those must be the newest jokes CCP came up with . Seriously, we've been playing long enough to know that all the so called maintenance you're doing during DT is just bull****. EvE server did have huge problems running the 23h a day its supposed to be able to do not long ago. And if am not mistaken most of the many unexpected downtimes were caused by database problems. So many other systems in the world have alot bigger databases than the one EvE uses, and they dont need dayly downtimes. The problem isn't the size of the database but the code your server is running. The code you devs are responsible for. Do you have a degree in Database management? Do you have a few years experience in dealing with SQL or other language? Seriously, as a developer for business technology, what you said demonstrates you know exactly -nothing- about database. I do get to touch database with headquarter going around 50 to 100 stores across the nation. Fact of the matter is that they DO requires daily maintenance and they are all painfully slow compare to eve. If you know a database sized remotely close to Eve that don't requires daily maintenance and / or runs faster then eve does. Please let me know which one it is cause I would love to hear more about those "So many other systems in the world have alot bigger databases than the one EvE uses, and they dont need dayly downtimes." you know off. Basically what you said sound like you are just pulling stuff out of your ass.
Ix Forres Vanguard Frontiers Imperial Republic Of the North	Posted - 2007.11.25 01:05:00 - [18] Originally by: SoldierOfJustice The problem isn't the size of the database but the code your server is running. The code you devs are responsible for. No, you're wrong. The database is huge. Google doesn't have daily downtime, but neither does Google require 100% concurrency for all clients, or 100% database consistency. Google shards, and Google has multiple clusters- in fact, Google does indeed have daily downtime- but you don't notice it because an alternative copy of the database is used for you while they update the other servers. CCP simply can't do that with the game. Downtime is here to stay, and who really cares? So what, you need to go outside for once in a while, shock horror... ISKsense \| Blog
Loedem Minmatar Knockaround Guys	Posted - 2007.11.25 01:07:00 - [19] Edited by: Loedem on 25/11/2007 01:08:07 Re: about google DT. Completely different type of DB there....with Google you're allowed to create redundancy. Because Eve is an unsharded persistant world, redundant DBs are not really an option...it'd be like having to constantly sync another dimension :P Now, if you were comfortable with server sharding, then maybe you could get your wish (I thought so) edit: ^^ hah, like minds...what he said :P
SoldierOfJustice Infortunatus Eventus HUZZAH FEDERATION	Posted - 2007.11.25 03:36:00 - [20] Originally by: Eka Maladay Do you have a degree in Database management? Do you have a few years experience in dealing with SQL or other language? Seriously, as a developer for business technology, what you said demonstrates you know exactly -nothing- about database. I do get to touch database with headquarter going around 50 to 100 stores across the nation. Fact of the matter is that they DO requires daily maintenance and they are all painfully slow compare to eve. If you know a database sized remotely close to Eve that don't requires daily maintenance and / or runs faster then eve does. Please let me know which one it is cause I would love to hear more about those "So many other systems in the world have alot bigger databases than the one EvE uses, and they dont need dayly downtimes." you know off. Basically what you said sound like you are just pulling stuff out of your ass. You're simply too stupid to even understand what I said. I do know about databases, and I am educated in the field. I do know SQL, and other languages. Now what point are you trying to make? Want to know about other systems with huge databases? could it be banks? search engines? or maybe other mmorpg's? or even the systems that you yourself work with? Re-read my post stupid. The point I was making isnt about the size of the database or the maintenance required. Its about the fact that the devs are using it as an execuse, 'cause if there was real maintenance then EvE wouldnt have been this f*ed up would it? Too many unexpected DT caused by DB failures prove them wrong when they say they do maintain the DB. They chose the 1h DT cause they thought it was enough, so when things go wrong it tells us that either the 1h isnt enough, or that they are not doing enough to maintain the DB. Since they havent made the DT longer it must be them who dont fix things as they supposed to do. Even CCP admitted that their design is old, and that includes the database design. Since technology moved a step further than what the server is running with it means that optimization is required for the server to serve the purpose with the big numbers of players and the data needed. Since you're too stupid to see what I ment here is a more simple version: Originally by:** CCP Gangleri With a database as massive as TQ it does not make sense to risk faults or undue lag by trying to run large jobs while it is live. A 30-60 minutes downtime a day ensures more stability and that the jobs that need to be run don't interfere with anything else. Faults and lag exist in eve, which are problems most of the times caused by database issues as CCP admitted. To ensure stability the devs looked at the situation when they designed the game and put "enough" time for maintaining the DB everyday. But thats apearenly not enough cause faults and lag caused by issues with the DB still happen. Either the 1h isnt enough, or that they are not doing enough to maintain the DB as I mentioned earlier, and in both cases its their fault. Understood? If not, then go see a doctor. Originally by: Ix Forres No, you're wrong. ... Downtime is here to stay, and who really cares? So what, you need to go outside for once in a while, shock horror... You read up. And you dont know me well enough to have the smallest idea about how much I go outside .
Eka Maladay	Posted - 2007.11.25 04:06:00 - [21] Originally by: SoldierOfJustice A bunch of unrelated bull**** Anyway, let me repeat my request. "If you know a database sized remotely close to Eve that don't requires daily maintenance and / or runs faster then eve does. Please let me know which one it is cause I would love to hear more about those "So many other systems in the world have alot bigger databases than the one EvE uses, and they dont need dayly downtimes." you know off." Thanks, now stfu.
SoldierOfJustice Infortunatus Eventus HUZZAH FEDERATION	Posted - 2007.11.25 06:47:00 - [22] Originally by: Eka Maladay Anyway, let me repeat my request. "If you know a database sized remotely close to Eve that don't requires daily maintenance and / or runs faster then eve does. Please let me know which one it is cause I would love to hear more about those "So many other systems in the world have alot bigger databases than the one EvE uses, and they dont need dayly downtimes." you know off." Thanks, now stfu. I couldnt find facts on ones without dayly DT, BUT there is SLAC's database. 900 Terabytes compressed data with 2.7hours of dayly DT and acquires 1 terabyte a day afaik , which means that eve's database need to be comparable to that to be as superb as you make it look like .
Ramirez Dora The Dead Parrot Shoppe Inc. Brutally Clever Empire	Posted - 2007.11.25 10:50:00 - [23] Google does not(nor do any other search engine vendor with half a brain) use a relational database model for the search indexes. Google has its own index format which is a bit slower than could be because they allow for the removal and insertion of documents into the index. And they do use an offline view of the index which gets updated during crawls and is then switched over to. Not to mention Google run their own server farms with like what, 80000 servers world wide? Comparing the kind of database architecture used for search indexes against Eve's is comparing apples and oranges. The fact that CCP works closely together with both Microsoft and the hardware vendor to push the technologies further shows that they are one of the leading implementations of enterprise database technology. Oh and that SLAC database is used to store and process particle physics research, it is also not comparable to Eve's DB in any way.
Thommy	Posted - 2007.11.25 11:36:00 - [24] To answer the question "when is downtime needed". Dureing where software or software components (files, executable code, functions etc) get updated. This can be for a number of reasons includeing but not limited to: - Bugfixes - Software updates (think about windows OS updates which serve for better stability & security) - Backups - Hardware replaceing Bugfixes Bugfixes sometimes can be done dureing an "hot" operation (which means an update while the software is still running fully active). Often you will need to reboot specific services though and if your operating system is not designed / configured to handle this or the service is to much depending on alot of other processes or alot of other processes are depending on this service it is nearly impossible to do it without causing alot of problems which could affect alot of people for a longer time then 1 hour. Software updates Most operating systems are updated from time to time to allow for improvements, bugfixes and security patches. In this particular case depending on your operating system it can or cant be done without an reboot but this also depends on which part is getting updated. Backups While there are alot of tools and alot of special hardware like mirroring, ghosting, raid and whatever else sometimes you will need to take down or lower the load on an software architecture to allow for an safe and proper backup. Real backups cannot be made while for example an database is running live and is continuesly recieving state changes, updates and inserts because when the backup process is halfway trough your data will be out of sync (which cause a wide varity of other problems). Hardware replacement Nothing works forever. It can always happen that an specific hardware component fails and cannot be replaced while the machine(s) that are depending on it are still powered up. Compare it with replaceing electric cables while the electricity is still on (it would be wise not to do this incase you dont know that allready). Not only these issues arrise when you never take an system offline but there are alot more reasons why from time to time its needed to bring in an downtime. I have tried running an machine 24x7 and i have been able to keep it running for 286 days (and some hours + minutes) but in the end i had to power down because the CPU fans started to act up. It survived 3 power failures thanks to an UPS, it has hot dockable harddisks with raid but in the end it came down to 2 simple fans that stopped working and couldnt be replaced without powering down (well unless i would take external measures to augment these specific fans). Trying to keep running would result in more and more external measures to try and augment failed components but it will eventually come down to an forced power down either because of total failure or the inability to sufficiently "patch up" the machine itself. When looking at an database itself and you try to stuff as much as possible into it while keeping it loaded and you perform heavy tasks on it while continueing with many other processes you will get an backlog. If this backlog becomes to large there is an real chance that the database itself crashes because the load increases by minutes of workload per second. Not only that but afterwards trying to fix issues like that will take alot longer then just not being impatient and giveing your database the time it needs to do its thing while not continueing to put heavy load on it. Guide \| Patch day
Jurgen Cartis Caldari Interstellar Corporation of Exploration	Posted - 2007.11.25 15:43:00 - [25] Originally by: SoldierOfJustice Originally by: Eka Maladay <Ad hominems> <More Ad Hominems> Originally by: CCP Gangleri With a database as massive as TQ it does not make sense to risk faults or undue lag by trying to run large jobs while it is live. A 30-60 minutes downtime a day ensures more stability and that the jobs that need to be run don't interfere with anything else. <Argument that if ANY lag and faults exist, CCP is doing nothing during DT> <Blaming everything on CCP devs> And, assuming you have programmed before, you should know that getting code perfect rarely happens. Particularly on large projects. Eve is not perfect. The Devs aren't gods, but you aren't either. If you've never ****** up code before, never gotten a segfault writing C, never a null pointer error, never accidentally created an infinite loop, you're a very special programmer. If not, get off your high horse and join the rest of us mortals down here. -------------------- ICE Blueprint Sales FIRST!! -Yipsilanti Pfft. Never such a thing as a "last chance". ;) -Rauth
Akane Miyamoto Paisti	Posted - 2007.11.25 16:08:00 - [26] Originally by: Jurgen Cartis If you've never ****** up code before, never gotten a segfault writing C, never a null pointer error, never accidentally created an infinite loop, you're a very special programmer. If not, get off your high horse and join the rest of us mortals down here. Amen. --------------- The Miyamotos, Chaper #1
SoldierOfJustice Infortunatus Eventus HUZZAH FEDERATION	Posted - 2007.11.25 16:57:00 - [27] Edited by: SoldierOfJustice on 25/11/2007 17:00:24 Originally by: Ramirez Dora Oh and that SLAC database is used to store and process particle physics research, it is also not comparable to Eve's DB in any way. Both eve and SLAC databases are required to work in realtime. Just because they use different technologies it doesnt mean they are incomparable.. Originally by: Jurgen Cartis And, assuming you have programmed before, you should know that getting code perfect rarely happens. Particularly on large projects. Eve is not perfect. The Devs aren't gods, but you aren't either. If you've never **** up code before, never gotten a segfault writing C, never a null pointer error, never accidentally created an infinite loop, you're a very special programmer. If not, get off your high horse and join the rest of us mortals down here. I've never said I am God, nor are the devs or you. Thats exactly why humans do make mistakes. We're not perfect. But if you keep making the same mistake over and over then there is something wrong with you. When you sell a product and try to maintain it you consider the tasks it performs and the amount of work needed to maintain it. If you do up like CCP does all the time, then you'll need more maintenance. Its not avoidable to ** up, but when you do you have to fix it asap and make sure your product works as intended. CCP intends to run the server 23/7. And they think only 7h are needed to maintain it everyweek. If that maintenance time isnt enough (and it obviously isnt and that was proven by our dear many unexpected DT) then THEY should do something to make their product work. CCP chose the requirements for EVE, not I, and not you. They think they can make it run, and it's obviously not running as well as intended, so it is their fault, they are the ones to blame. Not the aliens from outter space. Why do YOU and others get so angry when customers dont like what the service provider is giving them for their money? Do you work for CCP? If not then keep fixing your infinite loops to make sure your customers like your work. And before anyone starts telling me to do something else than play eve, remember that there are those of us who dont have much time to play, and when we do we are faced by stupid lag and/or unexpected DT. Edit: oh btw, your post is as much Ad hominems as mine and Eka Maladay's.

CCP Atropos	Posted - 2007.11.25 18:04:00 - [28] The TQ db may not be the largest in terms of entries in it, but in terms of transactions per day, it's greater than many international banks. In addition to this, there are many processes that we will only run at downtime; for example, converting outposts construction platforms into stations. Now, I'm not one of the DB Admins, so I can't comment on the how and the why of the daily downtimes, but I do know that without it, life would be that much more difficult for everyone involved.

Eka Maladay	Posted - 2007.11.25 20:05:00 - [29] Originally by: SoldierOfJustice Another bunch of bullcrap I happens to have played with SLAC before. While it probably is bigger, it is -no where- near as fast as Eve pulling query. Beside, it requires an even bigger downtime. So, what's your point? Other database either suck, or need DT just like eve do? I have no problem with you until you started to attack me and call me stupid. Mind you, but you are just pulling more horse crap out of whatever ass you have.
SoldierOfJustice Infortunatus Eventus HUZZAH FEDERATION	Posted - 2007.11.25 20:35:00 - [30] Originally by: Eka Maladay I happens to have played with SLAC before. While it probably is bigger, it is -no where- near as fast as Eve pulling query. Beside, it requires an even bigger downtime. So, what's your point? Other database either suck, or need DT just like eve do? I have no problem with you until you started to attack me and call me stupid. Mind you, but you are just pulling more horse crap out of whatever ass you have. You started it with this Originally by: Eka Maladay ... Basically what you said sound like you are just pulling stuff out of your ass. and you are too stupid to understand what I said. If you use time to fix something and when you're done its still broken then obviously you fixed nothing and there was no point in working on it at all. Who cares if you say you worked with SLAC? Do I believe you or those who designed it? You have no proof that you did work on any database at all. So you are pulling **** out of your ass. If I attacked CCP's work why do YOU care? Did I point a gun at YOUR head? Did I attack you in the beginning? I stated the fact that they arent doing enough to fix the stupid thing and you took it like I invaded your country. Cant believe the number of stupid players defending CCP when they have been in game less time, and donno half the problems CCP had trying to run EVE. If you are so facinated by EVE's DB and if you are as great at what you say you do then maybe you should work for CCP. Maybe then EVE wouldnt have problems caused by the DB.

Pages: [1] 2 3 4 :: one page
First page \| Previous page \| Next page \| Last page

COPYRIGHT NOTICE
EVE Online, the EVE logo, EVE and all associated logos and designs are the intellectual property of CCP hf. All artwork, screenshots, characters, vehicles, storylines, world facts or other recognizable features of the intellectual property relating to these trademarks are likewise the intellectual property of CCP hf. EVE Online and the EVE logo are the registered trademarks of CCP hf. All rights are reserved worldwide. All other trademarks are the property of their respective owners. CCP hf. has granted permission to EVE-Search.com to use EVE Online and all associated logos and designs for promotional and information purposes on its website but does not endorse, and is not in any way affiliated with, EVE-Search.com. CCP is in no way responsible for the content on or functioning of this website, nor can it be liable for any damage arising from the use of this website.