Best Linux Server uptime?

[Paperlantern]

Just wondering what the best server Uptime youve ever had with a Linux box (or really any server if you want to include other numbers as well). I have a Minecraft server thats running up on almost 90 days (its at 87 days and change right now) with no issues. I was thinking if my little crappy Minecraft box can pull off 3 months and running, i'm sure somebody with a linux box in production somewhere on nice hardware must have some awesome uptimes. Let's hear em!

Mine:

Cent OS 5.4 - Primary Function: Public Minecraft server for me and some friends. Uptime 87 days 16 hours 25 minutes.

 

[postaled]

Well my laptop at home is at around 115 days.... obviously I don't have automatic updates turned on, but I am looking to see how long I can go on it. Its running Ubuntu.


Also, does the Minecraft server run well now? Last time I tried blocks would not always get destroyed properly.

 

[Crusty]

09:46:19 up 286 days, 17:28, 1 user, load average: 0.02, 0.01, 0.00

Last reboot was due to hardware failure.

 

[Paperlantern]

Well my laptop at home is at around 115 days.... obviously I don't have automatic updates turned on, but I am looking to see how long I can go on it. Its running Ubuntu.


Also, does the Minecraft server run well now? Last time I tried blocks would not always get destroyed properly.

Yes, its out of Alpha now and is in Beta, there have been a butt ton of extra additions and fixes in even just the last couple months.

09:46:19 up 286 days, 17:28, 1 user, load average: 0.02, 0.01, 0.00

Last reboot was due to hardware failure.

Very nice, what OS? Cent OS? Red Hat? Some other distro? What do you use it for?

 

[zokudu]

Well my laptop at home is at around 115 days.... obviously I don't have automatic updates turned on, but I am looking to see how long I can go on it. Its running Ubuntu.


Also, does the Minecraft server run well now? Last time I tried blocks would not always get destroyed properly.

You sure that wasn't the spawn protection. The first 100 blocks around spoawn cannot be destroyed to prevent people griefing people who spawn.

 

[Paperlantern]

You sure that wasn't the spawn protection. The first 100 blocks around spoawn cannot be destroyed to prevent people griefing people who spawn.

This is only the case on certain servers running specific mods. A vanilla server like mine (though i'm looking at some mods now), if the player joins the server, they can play just like any other player. Thats why mine is publicly accessible, so my friends can get to it from their homes, but its not advertised in any way. You need mods to limit random players from griefing.

What I think he means is some time ago there was a bug where if you destroyed a block... say a cobble block with a diamond pick (or really any pick), you would see it disappear and it would do the destruction animation, but then another block would just pop back into existence. Very annoying. That HAS been fixed.

 

[postaled]

What I think he means is some time ago there was a bug where if you destroyed a block... say a cobble block with a diamond pick (or really any pick), you would see it disappear and it would do the destruction animation, but then another block would just pop back into existence. Very annoying. That HAS been fixed.

Yep. I was so excited but after trying that I just said screw multiplayer.

 

[Scarpozzi]

It wasn't uncommon for our servers to run over 365 days easy. I remember seeing uptime in the 400-550 day range for quite a few Red Hat 3 servers I used to run. The more common 'long running' servers were Netware 4.11/5.1 because they typically didn't run any 3rd party apps at that time. When other programmers wrote code for Netware other than Novell Engineers, they would typically get hosed within 6-8 months due to poor memory management. (memory space mgmt is a little different in Netware than Linux)

These days, kernel patches and SAN/infrastructure maintenance encourage scheduled shutdowns a few times a year. As long as you have proper UPS/generator to cover you in power outages, enough system memory to support future needs, and hot-swappable drives in a RAID array, the sky's the limit on uptime.

 

[Paperlantern]

Yep. I was so excited but after trying that I just said screw multiplayer.

Multiplayer has evolved a great deal in recent weeks. Netherworld is available, we have portals all over our server now to hop around the world quickly. Hes made accelerator track for carts so you dont have to MAKE them. The world has WEATHER now, it will suddenly cloud up and start pouring. It really is getting good. Every week theres something tweaked and improved or added.

An damn! A year plus! I had a feeling it was possible, though i suppose you are right... a little less possible now than in years past, but yeah, a good UPS system and RAID implementation and you could go for a long time.

 

[ultimatebob]

Most of my Linux systems only get rebooted for kernel patches every 180 days or so. If they need to get rebooted earlier than that, it's usually due to a power outage.

 

[Nothinman]

22:56:31 up 333 days, 2:34, 32 users, load average: 1.25, 1.28, 1.10

That's my MacPro running Debian sid at work, obviously I'm pretty far behind on the kernel updates. =)

 

[Khyron320]

We had some rehl 3 servers with over 1200 uptime. We had to rekick them though for security. At home I have a fedora 13 server with 152 day uptime with load averaging 1 at any given time.

lsof | grep deleted
great command on high uptime boxes

 

[sourceninja]

We do monthly reboots as part of our maintenance cycle for all servers.

 

[Lithium381]

We do monthly reboots as part of our maintenance cycle for all servers.

I've seen some boxes with uptime of 4+ years.....rediculous. If mine didn't suck so much juice i'd leave it on, but i shut it down when i don't need her.

 

[Lithium381]

22:56:31 up 333 days, 2:34, 32 users, load average: 1.25, 1.28, 1.10

That's my MacPro running Debian sid at work, obviously I'm pretty far behind on the kernel updates. =)

1.25 load? What are you running to slam it ?

 

[Nothinman]

1.25 load? What are you running to slam it ?

A 1.25 load isn't really that high and it's got 4 cores, 2 dual core Xeons I think, so I wouldn't say I'm slamming it. There's 3 or 4 VMs running under VMware Server in the background that are probably causing the spikes in CPU time, I don't even notice them though.

 

[sourceninja]

I've seen some boxes with uptime of 4+ years.....rediculous. If mine didn't suck so much juice i'd leave it on, but i shut it down when i don't need her.

It's not that they need it. It's to make sure they come back up on their own if they ever do get turned off. Patches and things always run the risk of making boot problems. Id rather find out when I have scheduled downtime.

 

[Modelworks]

I've got a pfsense box running on a p3-933 with 256MB ram that has been running nonstop for about 2 years now, no reboots or anything. I almost forget its there.

 

[Rifter]

my home file/game/print/firewall server has been running for almost a year. Before that it was up for almost 2 years. I shut it down about a year ago to kill off the RAID 5 array and switch up the storage configuration and upgrade the OS. Its running ubuntu.

 

[skyking]

443 days. We were moving it to a new building, and I toyed with the idea of moving it across on a cart hooked to the UPS. Turns out the UPS battery was getting weak.
Edit: I looked around and found one with 454 days

 

[Paperlantern]

HAHA that would be awesome, running the server accross on life support. Shoulda done it with some long white coats on, a video of that would have been hilarious. You guys barreling down a hallway, the UPS screaming its head off! if the box has redundant power supplies you could even run it off of two UPS units to extend the life. Yes. this would have been very cool to attempt.

454 days is awesome though.

 

[Rifter]

HAHA that would be awesome, running the server accross on life support. Shoulda done it with some long white coats on, a video of that would have been hilarious. You guys barreling down a hallway, the UPS screaming its head off! if the box has redundant power supplies you could even run it off of two UPS units to extend the life. Yes. this would have been very cool to attempt.

454 days is awesome though.

I did that to my server during a power outage, turned off all my other electronics and daisy chaned the UPS's together to keep the server running, good thing it was only a hour and a half outage.

 

[skyking]

HAHA that would be awesome, running the server accross on life support. Shoulda done it with some long white coats on, a video of that would have been hilarious. You guys barreling down a hallway, the UPS screaming its head off! if the box has redundant power supplies you could even run it off of two UPS units to extend the life. Yes. this would have been very cool to attempt.

454 days is awesome though.

I saw the battery was bad, I could have "hot swapped" in new battery but I did not have it on hand. I was tempted to go get one...............

 

[Scarpozzi]

I always heard this story, but never knew where it originated from. It isn't Linux, but Netware has a lot of similarities to Linux...even NW3 & NW4.:

http://arstechnica.com/civis/viewtopic.php?f=2&t=1824

 

[kedlav]

$uptime
2:42pm up 1103 day(s), 1:16, 2 users, load average: 0.30, 0.44, 0.51

That's a preprod Solaris box.

$uptime
14:43:50 up 279 days, 21:48, 4 users, load average: 7.14, 6.51, 6.16

That's our RHEL5 monitoring server

I could probably dig up a few more dev boxen that are even higher...