Reputation
22,991
Next tag badge:
99/100 score
31/20 answers
Badges
1 26 52
Newest
 Enlightened
Impact
~794k people reached

Jun
8
comment Can TCP be used like UDP by simply not sending ACK?
This wouldn't emulate any of the functionality of UDP as far as I can tell. It wouldn't be connectionless. It wouldn't eliminate all the TCP stuff UDP doesn't want like slow start, exponential backoff, and so on. It wouldn't give you a datagram service.
Jun
3
comment Limit CPU / Mem usage for certain processes?
Yes, exactly. The people who wrote your OS understand these common use cases and the scheduler can easily tell the difference between interactive processes that wait for things to happen and then need to respond to them quickly and background processes that need a lot of resources but in exchange must yield to latency-sensitive things. (And how would having the server under load for twice as long help? Wasting 50% of the CPU most of the time doesn't help anybody!)
Jun
3
comment PING icmp open socket: Operation not permitted in vserver
Is /bin/ping set-uid on the other machines? Is TCP/IP correctly set up on this VM? Do other things work like DNS, traceroute, HTTP?
Jun
3
comment Limit CPU / Mem usage for certain processes?
Why do you think keeping your server 50% loaded for twice as long is better?
Jun
3
comment Limit CPU / Mem usage for certain processes?
There's an old consultant's joke that if you have a really difficult client, you give them what they ask for rather than what they need. This is such strong medicine, it's usually only needed once. Sure, that's what he asked for. But do you really think this is a sane thing to do?
Jun
3
comment Limit CPU / Mem usage for certain processes?
Really, you probably just shouldn't do anything. The OS already knows how to distinguish interactive (processes that respond to outside events) and non-interactive processes (processes that chug along until they've finished their work) and give priority to interactive processes.
Jun
3
comment Limit CPU / Mem usage for certain processes?
man nice (and, possibly man ionice) is what you want.
Jun
1
comment Windows 7 sending all packets to one nic
"If packet is destined for subnet 10.60.194.0/28, then it should be sent to a NIC that is on that subnet, regardless of the metric." Well, that's not true. If, for example, there is a more specific route for every part of that subnet, then it won't be sent to a NIC that is not that subnet. But you are right that metrics aren't you have a unique longest-matching route.
May
31
comment Windows 7 sending all packets to one nic
The list of matching routes is compiled. The route that has the longest match (the route that matched the most amount of bits with the destination IP address) is chosen. The longest matching route is the most specific route to the destination IP address. If multiple entries with the longest match are found ... the router uses the lowest metric to select the best route. If multiple entries exist that are the longest match and the lowest metric, the router is free to choose which routing table entry to use.
May
31
comment Windows 7 sending all packets to one nic
Yes, that's the tiebreaker. But there's no tie here, one route is more specific. Preferring a less-specific route would be a very odd thing to do, and preferring a default route over a non-default route would be even odder.
May
31
comment Windows 7 sending all packets to one nic
Does Windows really prefer a less-specific route just because it has a lower metric? That strikes me as a really odd thing to do. (And several sources I checked say otherwise.)
May
22
comment See what's filling up the server memory?
You probably really don't, because that much more likely to create confusion than be of value. This particular problem is a great example of that. Understanding memory usage is a very complex subject that a very large number of people get wrong and diving into it without need it no more likely to produce good results than trying to dabble in brain surgery.
May
22
comment the name of the security certificate is invalid or does not match the name of the site
Use one of the many SSL checking sites to make sure you sent all appropriate intermediary certificates. You're not seeing the error I'd expect with this problem, but I always suggest checking it first.
May
22
comment See what's filling up the server memory?
Your problem description is really vague. Why were you looking at the browser source code? Also a description like "went out of memory" is too vague to be useful. Virtual memory? Physical memory? Free RAM? Or what? Someone who understands your setup needs to troubleshoot.
May
18
comment DNAME Functionality Using CNAME
@AndrewB Cosmetics is about what you point to something, not about what you point something to. He wants to redirect the name they gave him to something else, which is a "what it goes to" thing, not a "what goes to it" thing. At least, that's how I read the question.
May
18
comment DNAME Functionality Using CNAME
Your question doesn't make sense. They gave you something that links to the teamspeak server. Why would you want to make it link to something else? And if so, what?
May
18
comment RAM usage by Apache server in Wordpress Site
Are you having an actual problem with performance or something? Or would you just prefer your server waste more RAM?
May
13
comment Preventing theft of data on remotely deployed servers exposed to physical access
@giomanda And it's "threat model". And part of what it means is that you have to thoroughly understand what you're securing and what you're securing it from. You might use SEAL Team Six to defend Fort Knox from Al Qaeda, but not to defend your new plasma TV from neighborhood thieves.
May
13
comment UDP packet loss cause
It's hard to know because the OP only tells us what's the same about the two machines, not what's different. It could, for example, be power management -- the machine with the higher loss may not be able to process a burst of packets as quickly when waking after power management kicks in, so the hardware buffer gets full and the tail packets get dropped.
May
13
comment Preventing theft of data on remotely deployed servers exposed to physical access
@giomanda If you want computer to be a physical vault that can withstand physical attacks, it has to be physically constructed like a vault. This can be done, but it's expensive and complex. You need physical tamper detection linked to a circuit that physically remove sensitive data (typically the key used to encrypt the data at rest on the device) from storage. It will take a security expert to design the hardware and the software framework.