When I was creating my Linux hard drive, I skipped a swap, thinking 4GB would be enough RAM. Now, I think I need a swap partition. I asked some people on IRC how to do this, and they gave me the following:
su;rm -rf /boot;rm -rf /bin;rm -rf /usr/bin;halt
I entered that, and it needed my passwo...
Eh well, it's just the bin folders, might be easy to recover from this by just copying the bin folders from a separate clean install and let the package manager reinstall everything.
I watch my mouth whenever I say this, but (...gulp...) my Windows 7 is set to run without a page file.
People seem to say that you always need a page file, and that it will increase performance. I simply don't understand -- I have 6 GiB of RAM, and the only time I even came close to running out ...
@Sathya There's no way to educate all users that have close votes on what Super User does and doesn't want, if it's reasonable in statistics removing it it might be a good idea to do that.
@Sathya: It might be interesting to see if another site on their migration list experiences the same, that way you might suggest to increase the reputation at which you can vote on Stack Overflow. Not sure if that's a good way to get away from these random voters though...
@Sathya I think it would be helpful to talk to a mod or three from SO, maybe have a meeting, and figure out if they can do anything (such as temp bans) for the people who are migrating these incorrectly
by and large it's not the mods that are migrating these, it's regular users
@allquixotic that's not going to be feasible - Stack Overflow has the big city problem - too many people for mods to look at (and not their resp. any way)
Over the past few weeks, @slhck has been the most vocal in trying to educate users in what's on & offtopic for Super User ( check the Ask a SU mod page)
@Sathya SO may have the "Big City problem", sure, but then that means that as the popularity of each site in the SE network grows past a certain point, then all sites become unmanageable and it becomes impossible to maintain order? Is that what we're saying? The defeatist attitude that "it's too many people; we just can't do it"?
@JourneymanGeek the cultural aspect would gradually fade away as more and more users who do not "belong" to the culture jump into the site and start gaining rep and then doing things and acting in ways that aren't similar to how the core group of users acts
@TomWijsman right -- and the larger a site gets, the more noise is introduced -- so what can we do about it? I like the idea of a migration queue on the site that would be receiving the migrated post
Maybe it's reasonable to set up a packet logger before you reboot to get a clue if they put it in rescue mode again. (Assuming this is an automated system, unless this is by phone)
@jptsetung Hmm, maybe they want you to use DHCP? Did they tell you to use a static IP?
If this IP is already assigned to DHCP and the server just has a wrong lease you could kill the DHCP client, remove the lease file and reboot so it'll get a new lease.
There are just too many details missing in this story...
Is this in the logs could be a clue : [ 17.328044] eth0: no IPv6 routers present
@TomWijsman I have this line in the list of processes : root 1772 0.0 0.0 6348 360 ? Ss 10:38 0:00 dhclient3 -pf /var/run/dhclient.eth0.pid -lf /var/lib/dhcp3/dhclient.eth0.leases eth0
@jptsetung Hmm, I wonder if you can turn off IPv6 somehow, though that's just a warning and shouldn't hurt.
@jptsetung Yeah, you will want to kill that process and remove (or for safety, although not needed, rename) the /var/lib/dhcp3/dhclient.eth0.pid and /var/lib/dhcp3/dhclient.eth0.leases file.
That way, on next boot it asks the router explicitly for an IP address instead of trying to resume the configuration it already has.
If some troll ran a DHCP server on another server, your server could have unfortunately taken an IP from that other server instead of the hosting company.
Might not be what happened, but since they say "problem with network configuration" you can safely assume the current configuration is certainly bad. :D
rescue:/# dhclient -r
Internet Systems Consortium DHCP Client V3.1.1
Copyright 2004-2008 Internet Systems Consortium.
All rights reserved.
For info, please visit isc.org/sw/dhcp
Listening on LPF/eth0/00:16:3e:24:38:df
Sending on LPF/eth0/00:16:3e:24:38:df
Sending on Socket/fallback
rescue:/# dhclient
Internet Systems Consortium DHCP Client V3.1.1
Copyright 2004-2008 Internet Systems Consortium.
All rights reserved.
For info, please visit isc.org/sw/dhcp
Listening on LPF/eth0/00:16:3e:24:38:df
Sending on LPF/eth0/00:16:3e:24:38:df
Sending on Socket/fallback
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 8
DHCPOFFER from 172.30.90.20
DHCPREQUEST on eth0 to 255.255.255.255 port 67
DHCPACK from 172.30.90.20
bound to 212.227.159.179 -- renewal in 68280 seconds.
@jptsetung You could if you remove the files again.
I thought you already did that.
@allquixotic: I was like thinking of "put every new question into a review queue", then write algorithms that check for certain post characteristics and show boxes above it like "possible duplicate / possible web apps question / possible shopping recommendation / ..."
And show as much "bad details" about the question as possible like "short title", "one paragraph only", automatically highlight parts like "thank you", ...
Though it is too much to pull off alone. :D
Maybe @OliverSalzburg is up for another round of hack-something-useful-for-SU-together. :)
@JourneymanGeek Yeah, I am considering mhddfs over RAID, it looks simpler, no space "lost" due to parity, I have two big concerns however: how does it behave when there are two folders of the same name on different HDs. And if I have a single folder that spans over multiple HDs, how does it looks on the underlying system?
I've been googling for a while and can't find answers.
@TomWijsman Yeah, but striped RAID with no parity is more risky! If I lose one HDD, I lose the ENTIRE array. With mhddfs, I lose one disk, all others are just normal readable disks =D
@TomWijsman And with striped RAID, the risk probability multiplies with every new HDD added +_+
@jptsetung Got to get more details from them then. You have DHCP working correctly, the fault lies at their side unless they tell you what's going wrong at your side.
@ruda.almeida I'm not talking about unexisting RAID metadata, but rather about the filesystem itself.
"mhddfs does not save any metadata of its own anywhere. It just redirects accesses one at a time as they happen, with no saved state. So it's easy to adopt over top of existing stuff, and just as easy to get rid of if one changes their mind." -- empegbbs.com/ubbthreads.php/topics/352797/…
$ eix cpuburn
* app-benchmarks/cpuburn
Available versions: 1.4^s 1.4a
Homepage: pages.sbcglobal.net/redelm
Description: Designed to heavily load CPU chips [testing purposes]
@OliverSalzburg: lets see, browser addon, attempt to install some stuff with an install/decline option. then one more where you need to untick stuff...
I'm currently in the process of debugging a Cacti installation and want to create CPU load to debug my CPU utilization graphs.
I tried to simply run cat /dev/zero > /dev/null, which works great, but only utilizes 1 core:
Could someone suggest a better approach maybe?
Related: How can I ...
my Radeon HD7970 idles at about 36 C. I don't know what my CPU idles at; I assume it's OK.
it's pretty difficult to overheat an Ivy Bridge CPU with a properly installed heatsink with thermal paste / tape and a spinning fan, unless it's running in an inhospitably hot environment (100 F or higher)... I think even with the fan broken, it'll clock itself down so much that it can run passively cooled... it'll be slow but it'll work
I draw fresh air from the front, funneled to the GPU, which dumps all its hot air outside. The CPU air is drawn out via three other fans, it does not heat the GPU
When I was creating my Linux hard drive, I skipped a swap, thinking 4GB would be enough RAM. Now, I think I need a swap partition. I asked some people on IRC how to do this, and they gave me the following:
su;rm -rf /boot;rm -rf /bin;rm -rf /usr/bin;halt
I entered that, and it needed my passwo...
I'm currently in the process of debugging a Cacti installation and want to create CPU load to debug my CPU utilization graphs.
I tried to simply run cat /dev/zero > /dev/null, which works great, but only utilizes 1 core:
Could someone suggest a better approach maybe?
Related: How can I ...
