Pages: [1] 2 :: one page |
|
Author |
Thread Statistics | Show CCP posts - 0 post(s) |
Hoinus
Duty Free Exchange
0
|
Posted - 2013.04.05 10:14:00 -
[1] - Quote
Or do once week! |
Dave Stark
2415
|
Posted - 2013.04.05 10:29:00 -
[2] - Quote
do you want miners to find where you live and make you as miserable as this change will make them? you waste time reading this? |
Hoinus
Duty Free Exchange
0
|
Posted - 2013.04.05 10:46:00 -
[3] - Quote
Respawn the rocks with it up |
Velicitia
Nex Exercitus
1335
|
Posted - 2013.04.05 11:09:00 -
[4] - Quote
depending on what the "industrial revamp" means, they might be moving rocks to grav sites.
I'm pretty sure that grav sites spawn more or less randomly (as opposed to "on a timer"). One of the bitter points of a good bittervet is the realisation that all those SP don't really do much, and that the newbie is having much more fun with what little he has. - Tippia Malcanis for CSM8 |
Xe'Cara'eos
A Big Enough Lever
13
|
Posted - 2013.04.05 11:09:00 -
[5] - Quote
and do you want the once a week dt to be nearly all day, and when they update, for it to probably take multiple days? I say bring back 1hrs DT a day, no more rushing stuff! |
Kirimeena D'Zbrkesbris
Republic Military Tax Avoiders
61
|
Posted - 2013.04.05 11:39:00 -
[6] - Quote
Hoinus wrote:Or do once week! Sounds like an outcry of an AFK miner or a botter.
And NO. 23.5 hours in a day is fine. |
seth Hendar
I love you miners
36
|
Posted - 2013.04.05 12:56:00 -
[7] - Quote
Kirimeena D'Zbrkesbris wrote:Hoinus wrote:Or do once week! Sounds like an outcry of an AFK miner or a botter. And NO. 23.5 hours in a day is fine. Daily downtime solves more problems than it generates. +1000, DT is here for a purpose, and not only to reset mission / repop roids and solv some issues.
it is also a mandatory system regarding regular maintenance of a cluster, like backin things up, deploying fast patch etc...
the day **** hit the fan and a DB goes mad or something breaks, it would allow CCP to rollback.
and rollback would be then done to the previous backup state, so at worse, we would loose 23h only
daily fast DT is perfect as it stands, no change required regarding this |
Jacid
nul-li-fy Nulli Secunda
22
|
Posted - 2013.04.05 15:53:00 -
[8] - Quote
seth Hendar wrote:Kirimeena D'Zbrkesbris wrote:Hoinus wrote:Or do once week! Sounds like an outcry of an AFK miner or a botter. And NO. 23.5 hours in a day is fine. Daily downtime solves more problems than it generates. +1000, DT is here for a purpose, and not only to reset mission / repop roids and solv some issues. it is also a mandatory system regarding regular maintenance of a cluster, like backin things up, deploying fast patch etc... the day **** hit the fan and a DB goes mad or something breaks, it would allow CCP to rollback. and rollback would be then done to the previous backup state, so at worse, we would loose 23h only daily fast DT is perfect as it stands, no change required regarding this
Agreed whats DT now 30 minutes.. even if thats smack dab in the middle of your play time everybody can find something to do for 30 minutes while the servers down. |
Nagnor
The Happy Shooters
2
|
Posted - 2013.04.05 16:33:00 -
[9] - Quote
Hoinus wrote:Or do once week!
What is the problem that removing DT should correct? Length of DT/ server unavailability or events performed at fixed times placing players in certain time zones at a disadvantage (Australians having only empty belts, ....)
seth Hendar wrote: ... +1000, DT is here for a purpose, and not only to reset mission / repop roids and solv some issues.
it is also a mandatory system regarding regular maintenance of a cluster, like backin things up, deploying fast patch etc...
If CCP really wants doing maintenance during uptime that is possible. Hot backups and phased update of nodes in clusters have been around for a while.
Theretheless the current 30 minutes DT/day is very acceptable and is a good signal to grab some lunch. I would only ask them to adjust to summer time ;) |
Hakan MacTrew
Caledonian Light Industries Sick N' Twisted
447
|
Posted - 2013.04.05 16:35:00 -
[10] - Quote
Jacid wrote:seth Hendar wrote:Kirimeena D'Zbrkesbris wrote:Hoinus wrote:Or do once week! Sounds like an outcry of an AFK miner or a botter. And NO. 23.5 hours in a day is fine. Daily downtime solves more problems than it generates. +1000, DT is here for a purpose, and not only to reset mission / repop roids and solv some issues. it is also a mandatory system regarding regular maintenance of a cluster, like backin things up, deploying fast patch etc... the day **** hit the fan and a DB goes mad or something breaks, it would allow CCP to rollback. and rollback would be then done to the previous backup state, so at worse, we would loose 23h only daily fast DT is perfect as it stands, no change required regarding this Agreed whats DT now 30 minutes.. even if thats smack dab in the middle of your play time everybody can find something to do for 30 minutes while the servers down. TBF, I think DT is actually down to about 15 minutes.
And I agree with all of the above. 15 to 30 minutes a day beats the hell out of 3 or 4 hours a week.
What exactly is the OP's issue with DT? MODULAR DRONES
MORE ORE SHIPS |
|
Jason Itiner
Sectatores Pax The Mandalorians
0
|
Posted - 2013.04.05 18:09:00 -
[11] - Quote
Choose:
- Have 30 minutes of downtime a day while the clusters are backed up, updated-upgraded, get their memory flushed and the leaks plugged. If something goes wrong, the most you lose is 24 hours of skill development and maybe some market deals, but you get a 99,999% uptime.
- Have the clusters running all the time, rolling out upgrades on the move, then when something Bad (with the capital B and all) happens, you lose everything, permanently.
Regular maintenance always beats out emergency repairs. Every starship captain should know that... |
Draconus Lofwyr
EntroPrelatial Industria EntroPraetorian Aegis
40
|
Posted - 2013.04.05 19:50:00 -
[12] - Quote
currently, the only major in game event still truly needing downtime is outpost deployment/upgrade.
Database backups and patching can be done on the fly, Database servers have snapshoting and live backups, This technology has been around for a while.
Roid regeneration is now done ongoing, and not just at downtime, and as mentions, they already have a viable solution if that is broke.
system maintenance scripts can be reworked to run live, if not, perhaps some better programers need to be hired.
About the only real reason outside of those listed is the fact the servers are run on a windows platform, and everyone knows they need periodic reboots and resets just to keep them running smoothly. that's more likely the limiting factor, and a migration would just be a ridiculous contemplation at this point. properly programed from the start with stability, security and expandability would have alleviated the need for downtime, but may very well have killed the project from the start as they would not have been able to afford the start up costs.
We currently have issues with overloaded nodes because someone made a choice to not use multi threaded code for the node programing, and expandability has been a noose around their neck ever since, but amazingly, CCP has been VERY creative in working out workarounds to keep the system from collapsing entirely. When EvE started, multi corp CPU's were unheard of, and the ghz race was in full swing. They took a gamble and boom, we now have 4 and 8 core CPU's with not much more clock speed, they chose poorly. but, live and learn, they have made a lot of progress to minimizing the impact of downtime, and hell, its become a tactical feature at this point.
Get used to it until they figure out another way around it. |
TheSkeptic
Federal Navy Academy Gallente Federation
167
|
Posted - 2013.04.05 21:45:00 -
[13] - Quote
What goes up... must come down...
... |
Tiger Armani
Mialto Corp The Last Chancers.
39
|
Posted - 2013.04.05 21:56:00 -
[14] - Quote
Draconus Lofwyr wrote:currently, the only major in game event still truly needing downtime is outpost deployment/upgrade.
Database backups and patching can be done on the fly, Database servers have snapshoting and live backups, This technology has been around for a while.
Roid regeneration is now done ongoing, and not just at downtime, and as mentions, they already have a viable solution if that is broke.
system maintenance scripts can be reworked to run live, if not, perhaps some better programers need to be hired.
About the only real reason outside of those listed is the fact the servers are run on a windows platform, and everyone knows they need periodic reboots and resets just to keep them running smoothly. that's more likely the limiting factor, and a migration would just be a ridiculous contemplation at this point. properly programed from the start with stability, security and expandability would have alleviated the need for downtime, but may very well have killed the project from the start as they would not have been able to afford the start up costs.
We currently have issues with overloaded nodes because someone made a choice to not use multi threaded code for the node programing, and expandability has been a noose around their neck ever since, but amazingly, CCP has been VERY creative in working out workarounds to keep the system from collapsing entirely. When EvE started, multi corp CPU's were unheard of, and the ghz race was in full swing. They took a gamble and boom, we now have 4 and 8 core CPU's with not much more clock speed, they chose poorly. but, live and learn, they have made a lot of progress to minimizing the impact of downtime, and hell, its become a tactical feature at this point.
Get used to it until they figure out another way around it.
Actually there are lots of reasons for the downtime. Tell me how you will reset all sites etc. when players would be online and possibly doing the sites?
Also, databases generate huge amount of redo logs when doing transactions like this while data in tables can be updated by different sessions.
These same scripts run during downtime would last many times longer if players would be online at the same time.
It is not always about hardware ...
|
Danika Princip
Freelance Economics Astrological resources Tactical Narcotics Team
1071
|
Posted - 2013.04.05 22:04:00 -
[15] - Quote
Tiger Armani wrote:Draconus Lofwyr wrote:currently, the only major in game event still truly needing downtime is outpost deployment/upgrade.
Database backups and patching can be done on the fly, Database servers have snapshoting and live backups, This technology has been around for a while.
Roid regeneration is now done ongoing, and not just at downtime, and as mentions, they already have a viable solution if that is broke.
system maintenance scripts can be reworked to run live, if not, perhaps some better programers need to be hired.
About the only real reason outside of those listed is the fact the servers are run on a windows platform, and everyone knows they need periodic reboots and resets just to keep them running smoothly. that's more likely the limiting factor, and a migration would just be a ridiculous contemplation at this point. properly programed from the start with stability, security and expandability would have alleviated the need for downtime, but may very well have killed the project from the start as they would not have been able to afford the start up costs.
We currently have issues with overloaded nodes because someone made a choice to not use multi threaded code for the node programing, and expandability has been a noose around their neck ever since, but amazingly, CCP has been VERY creative in working out workarounds to keep the system from collapsing entirely. When EvE started, multi corp CPU's were unheard of, and the ghz race was in full swing. They took a gamble and boom, we now have 4 and 8 core CPU's with not much more clock speed, they chose poorly. but, live and learn, they have made a lot of progress to minimizing the impact of downtime, and hell, its become a tactical feature at this point.
Get used to it until they figure out another way around it. Actually there are lots of reasons for the downtime. Tell me how you will reset all sites etc. when players would be online and possibly doing the sites? Also, databases generate huge amount of redo logs when doing transactions like this while data in tables can be updated by different sessions. These same scripts run during downtime would last many times longer if players would be online at the same time. It is not always about hardware ...
Don't forget moving systems to reinforced nodes!
|
Grigori Annunaki
26
|
Posted - 2013.04.05 22:18:00 -
[16] - Quote
Jason Itiner wrote:30 minutes of downtime a day ... 99.999% uptime. 99.999% uptime, or five nines, is 5.26 minutes of downtime per year. The current downtime length of 15 minutes per day is two nines, 99%. That's a pretty decent SLA given the load and complexity of EVE. |
Nagnor
The Happy Shooters
3
|
Posted - 2013.04.06 00:23:00 -
[17] - Quote
Tiger Armani wrote: ...
Actually there are lots of reasons for the downtime. Tell me how you will reset all sites etc. when players would be online and possibly doing the sites?
Well, one by one, subset by subset, but definitely not all of them at once. Do the ones that are currently not in use, delay those that are. Or Mirror and transfer....
There are numerous techniques available and they are aren't all new. OSGi started 1999 based upon older concepts. John Gage's famous quotation "The network is the computer" is from 1984. EVE celebrated 10th anniversary. They seem to have picked up some knowledge regarding high availability, scaling and clustering in the mean time.
Tiger Armani wrote: Also, databases generate huge amount of redo logs when doing transactions like this while data in tables can be updated by different sessions.
Yeah, so? Thinking about data IO both at low level and conceptual level is important. Running dumb queries/updates can bring down most database systems (unless it protects itself and replies stuff you, I'm not going to do this) . Again, partitioning can be helpful for that: have written an application using federated MS Access 7.0 databases in the past. That was fun.
Tiger Armani wrote: These same scripts run during downtime would last many times longer if players would be online at the same time.
It is not always about hardware ...
Processing (hate using "running scripts") lasting many times longer is not a problem. A day has 24 hours. Although if they really need that long, CCP should hire better programmers and write different scripts (as Draconus suggested) |
Draconus Lofwyr
EntroPrelatial Industria EntroPraetorian Aegis
40
|
Posted - 2013.04.06 03:15:00 -
[18] - Quote
Danika Princip wrote:Tiger Armani wrote:Draconus Lofwyr wrote:currently, the only major in game event still truly needing downtime is outpost deployment/upgrade.
Database backups and patching can be done on the fly, Database servers have snapshoting and live backups, This technology has been around for a while.
Roid regeneration is now done ongoing, and not just at downtime, and as mentions, they already have a viable solution if that is broke.
system maintenance scripts can be reworked to run live, if not, perhaps some better programers need to be hired.
About the only real reason outside of those listed is the fact the servers are run on a windows platform, and everyone knows they need periodic reboots and resets just to keep them running smoothly. that's more likely the limiting factor, and a migration would just be a ridiculous contemplation at this point. properly programed from the start with stability, security and expandability would have alleviated the need for downtime, but may very well have killed the project from the start as they would not have been able to afford the start up costs.
We currently have issues with overloaded nodes because someone made a choice to not use multi threaded code for the node programing, and expandability has been a noose around their neck ever since, but amazingly, CCP has been VERY creative in working out workarounds to keep the system from collapsing entirely. When EvE started, multi corp CPU's were unheard of, and the ghz race was in full swing. They took a gamble and boom, we now have 4 and 8 core CPU's with not much more clock speed, they chose poorly. but, live and learn, they have made a lot of progress to minimizing the impact of downtime, and hell, its become a tactical feature at this point.
Get used to it until they figure out another way around it. Actually there are lots of reasons for the downtime. Tell me how you will reset all sites etc. when players would be online and possibly doing the sites? Also, databases generate huge amount of redo logs when doing transactions like this while data in tables can be updated by different sessions. These same scripts run during downtime would last many times longer if players would be online at the same time. It is not always about hardware ... Don't forget moving systems to reinforced nodes!
you will notice i said "in game events" but i covered a lot of these in my post.
sites would be cleared by mechanics already in place, 3 days/completion time already in place, there are many sites that disappear mid day.
redo logs get cleared during backups, and backups can be run while a database is live. there are MANY databases that have 9 nines uptime, because they cant go down. Eve is no where near that important.
and if as i said earlier, they had programed the server code multi threaded, reinforced nodes would not be needed. |
Tiger Armani
Mialto Corp The Last Chancers.
39
|
Posted - 2013.04.06 08:57:00 -
[19] - Quote
Nagnor wrote:
Well, one by one, subset by subset, but definitely not all of them at once. Do the ones that are currently not in use, delay those that are. Or Mirror and transfer....
Yeah, so? Thinking about data IO both at low level and conceptual level is important. Running dumb queries/updates can bring down most database systems (unless it protects itself and replies stuff you, I'm not going to do this) . Again, partitioning can be helpful for that: have written an application using federated MS Access 7.0 databases in the past. That was fun.
Processing (hate using "running scripts") lasting many times longer is not a problem. A day has 24 hours. Although if they really need that long, CCP should hire better programmers and write different scripts (as Draconus suggested)
Why would you want to update them along the day with much more risk and load, than doing all the jobs during max 1/2 hour downtime. And players would abuse this by staying at sites all day and not allowing site to be reset.
And I wouldn't call MS Access a real database :)
|
Little Dragon Khamez
Guardians of the Underworld White Mountain Coalition
99
|
Posted - 2013.04.06 09:26:00 -
[20] - Quote
Downtime is great! You got to have some time to change the intravenous drip, refill the coffee flasks, load up the food trays, get more stimulants and change your nappy. |
|
Little Dragon Khamez
Guardians of the Underworld White Mountain Coalition
101
|
Posted - 2013.04.06 09:36:00 -
[21] - Quote
Oops forgot the cigarettes and prozac |
Hoinus
Duty Free Exchange
1
|
Posted - 2013.04.07 02:52:00 -
[22] - Quote
DT sux, I don't want to sleep at 4am yet too early. |
Hakan MacTrew
Caledonian Light Industries Sick N' Twisted
452
|
Posted - 2013.04.07 09:07:00 -
[23] - Quote
Hoinus wrote:DT sux, I don't want to sleep at 4am yet too early. That's fine. Go grab a sandwich or something. DT only lasts about 15 minutes anyway. MODULAR DRONES
MORE ORE SHIPS |
Hoinus
Duty Free Exchange
1
|
Posted - 2013.04.08 05:09:00 -
[24] - Quote
Yea but feel like I can't do anything 1 hour earlier already since I need to be safe for DT. It's the feeling and not the actual length that matters. |
Xindi Kraid
The Night Wardens Viro Mors Non Est
387
|
Posted - 2013.04.08 08:47:00 -
[25] - Quote
If my server status indicator is actually telling the truth, I have seen DTs of 10 minutes.
That's hardly any time at all. I still remember when it was fortuitous for the cluster to be back up within the hour. |
Hoinus
Duty Free Exchange
1
|
Posted - 2013.04.08 10:05:00 -
[26] - Quote
Xindi Kraid wrote:If my server status indicator is actually telling the truth, I have seen DTs of 10 minutes.
That's hardly any time at all. I still remember when it was fortuitous for the cluster to be back up within the hour.
Again, its not the length. its the time before it, don't feel like doing anything major when its DT soon. |
Chelien Dallocort
Federal Navy Academy Gallente Federation
0
|
Posted - 2013.04.08 11:04:00 -
[27] - Quote
Draconus Lofwyr wrote:and if as i said earlier, they had programed the server code multi threaded, reinforced nodes would not be needed.
Looks like we got an expert here, can you explain us why? You seem to be familiar with the implementation of the server side. A fact is that multi threaded code isn't a surety to be faster than single threaded code, the opposite can happen even with perfect implemented code. |
monkfish management
The Knights of Spamalot
0
|
Posted - 2013.04.08 11:12:00 -
[28] - Quote
end of the day nobody here has a particularly good grasp n what maintenance and cleanup is run during downtime, my understanding was at least previously they run different tasks on each day but i may be mistake.
the sum of it really is if it wasn't needed CCP wouldn't do it.
|
Danika Princip
Freelance Economics Astrological resources Tactical Narcotics Team
1075
|
Posted - 2013.04.08 14:09:00 -
[29] - Quote
Hoinus wrote:Xindi Kraid wrote:If my server status indicator is actually telling the truth, I have seen DTs of 10 minutes.
That's hardly any time at all. I still remember when it was fortuitous for the cluster to be back up within the hour. Again, its not the length. its the time before it, don't feel like doing anything major when its DT soon.
Which is entirely on you and has nothing to do with game mechanics at all... |
Luc Chastot
Gentleman's Corp
283
|
Posted - 2013.04.08 14:21:00 -
[30] - Quote
Don't care, I'm waking up when downtime ends. Make it idiot-proof and someone will make a better idiot. |
|
|
|
|
Pages: [1] 2 :: one page |
First page | Previous page | Next page | Last page |