what about changing the domain names pointers for the thousands of domains pointing to the previous ip addresses? often in name servers not controlled by you? going through customers who, most of the time, dont even have a clue as to what a nameserver is?
i didn't say i had a solution, in fact i don't think there is one. on one side you have bloating of routing tables, on the other huge pains for anyone whose name is not MCI (or any big ISP)
policy problems
on
IPv6 is Here
·
· Score: 3, Insightful
an ip address for every human being... and they're non portable great!
i've said it before, and i'll say it again: ipv6 looks good on paper, but their current policy of not assigning IPs to anyone but big isps who will in turn sub delegate them to others is hindering the usefulness to small and medium ISPs
basicly you'll be locked into one isp, or face a major renumbering burden due to the non-portability of the addresses (and no it does NOT involve simply switching the network part)
yeah and im in security but listen to pretty much everything in there, i'll even watch britney for the eyecandy (well i did a few years ago when she first started)
on the other hand, i know a lot of headbanging developpers, and the one MS guy we have hear looks like the kind that would love britney...
how about the fact that you can't see the MAC address past the first hop? or the other that MAC addresses aren't (and don't need to be) garanteed to be globally unique?
alright. i'll rephrase that. the X server ISN'T being swapped out. better now?
and i'm using %s because it's just easier for a quick post on slashdot. if i ever want to do benchmarks i'll use hard numbers
and that still doesn't explain the 150 megs (happy now?) of memory use by firefox. that's RSS, not VSZ
enough of it. until you (or someone else) can give me a solution, or at least a reasonable explanation to that insane memory usage, i stand by what i said.
and quit spouting off your bullshit about caches, and read what i said properly. i already stated this isn't a cache issue, it's a pure memory issue
actually i don't use KDE, and haven't in over two years. i'm currently using openbox, and blackbox before that (moved due to lack of xinerama support)
right now i have a little over 20 applications loaded, of which a lot are terminals. the X server itself takes 10% of my available ram. firefox takes 15% (one window, one tab), openoffice 9% (with one 3 pages document loaded), the java vm 7% (proprietary CVS-like application) and evolution 7% (with lots of mails and accounts, that one is reasonable).
now that's some of the lowest numbers i've ever seen. it isn't rare the X process is around 40-50% along with firefox. that doesn't leave much for the rest of my applications to run, especially since the X process can't be swapped out (always being used).
this box is a p3 500 with 375 megs of ram. it IS insane use of memory. linux can run on a decade-old computer in shell mode no problem, but don't even think of running X on it (i tried, it "run" but, like running windows xp on a 486, you won't do much with it)
btw distro is debian sid, with the "prelink" package (did the same before i installed it too)
firefox on my system routinely use 50% of my total ram (out of 320). that's with only 1 window open and 5-10 tabs. i'd consider that "excessive memory use"
it also does the same when im at work
though linux (the whole deal, whichever distro you use, not just the kernel) in general tend to consume insane amounts of ram, firefox/mozilla is the worst so far
Does your Web site launch automatic download prompts? Microsoft will now suppress downloads not initiated by the user directly
holy shit! does that mean we will get sourceforge/download.com/cnet.com and all the other morons that make you jump through hoops to download a simple file to redeem their ways? naaaa unlikely
i run a small ISP, and i have the callerid of everyone calling, no matter what their privacy setting says. it even gets logged in my cute little radius database
as someone pointed out, it's a part of the ISDN call setup protocol.
except when your zones get around to morons ISP that force your zones to have a higher TTL than you specified, which unfortunately here is being done by the big player
this is a giant step FORWARD. if it can keep my network from being bombarded by all those damned windows viruses it's GOOD no matter what. and i don't even use windows.
i'd say this is the brightest idea microsoft had in the last decade (if they deliver that is)
there are tools to modify the bind config file too, that's not the point. your tools won't help one bit if you want to READ the damn file and understand what it does, unless maybe by converting it to BIND format which defeats the whole purpose (also the bind format just happen to be the dns protocol format)
the documentation might have been there, but it sure as hell wasn't clearly linked on the main page of djbdns, as i remember spending an hour or two looking for it on cr.yp.to
as for the multiple characters. i'm suggesting there is only one character to do something, and use two lines to do the two different things. in the end it's much clearer
i could care less to grep out all the A records. in the 4 years i've been here i never had a use for that. usually im looking for one well defined domain name (like the MX, mail.domain.com) or an IP address (when changing the IP of a server for exemple).
another toy that sound cool on paper but really isn't that useful
i have both a djbdns server (for a customer, 1200 domains or so) and a couple of bind ones (~400 domains).
how the fuck can you say djbdns is easier than bind? if i want an A record in bind it's "IN A" (see, easily understood). if you want the same in djbdns it's some cryptic characters that make no sense (and is, of course, undocumented, or was last time i needed it).
now the best part. there's MULTIPLE characters to do nearly the same thing. if i recall a + is a straight A record, and a @ (i think) is an A+PTR
give me bind anytime, it's MUCH easier. though i'm about to move to powerdns or something with a mysql or ldap backend so customers can edit their zones easily
that's right but their definition of ISP is limited. for them an ISP is an entity that will make sub allocations to other entities.
i currently work for an "ISP" that is too small to ever need to do sub allocations, but big enough that renumbering would involve about 2 months of work (with appropriate notifications to customers)
here's the appropriate entry in the policy:
5.1.1. Initial allocation criteria
a) be an LIR;
b) not be an end site;
c) plan to provide IPv6 connectivity to organizations to which it will assign/48s, by advertising that connectivity through its single aggregated address allocation; and
d) have a plan for making at least 200/48 assignments to other organizations within two years.
ipv6 has issues. mostly policy issues but that's pretty major IMHO
according to ARIN, under ipv6 you can no longer get allocated addresses for your own use. unless you're a big isp (tier 1) that plan to allocate addresses yourself (and lots of them).
that means the small isp that currently has a/20 (or lower) get to renumber all of their servers if they ever want to switch to another upstream provider, which means you're pretty much fucked if your current ISP goes under or tries to screw you over (yes it happened to me). trust me, renumbering involve more than just setting a new IP address on the server, it's a HUGE pain in the ass. the older your network is, the more of a pain it is
i understand the why of that policy (to prevent fragmentation), but it's still annoying
Slashdot Mirror
what about changing the domain names pointers for the thousands of domains pointing to the previous ip addresses? often in name servers not controlled by you? going through customers who, most of the time, dont even have a clue as to what a nameserver is?
i didn't say i had a solution, in fact i don't think there is one. on one side you have bloating of routing tables, on the other huge pains for anyone whose name is not MCI (or any big ISP)
an ip address for every human being... and they're non portable great!
i've said it before, and i'll say it again: ipv6 looks good on paper, but their current policy of not assigning IPs to anyone but big isps who will in turn sub delegate them to others is hindering the usefulness to small and medium ISPs
basicly you'll be locked into one isp, or face a major renumbering burden due to the non-portability of the addresses (and no it does NOT involve simply switching the network part)
yeah and im in security but listen to pretty much everything in there, i'll even watch britney for the eyecandy (well i did a few years ago when she first started)
on the other hand, i know a lot of headbanging developpers, and the one MS guy we have hear looks like the kind that would love britney...
download porn
how about the fact that you can't see the MAC address past the first hop? or the other that MAC addresses aren't (and don't need to be) garanteed to be globally unique?
alright. i'll rephrase that. the X server ISN'T being swapped out. better now?
and i'm using %s because it's just easier for a quick post on slashdot. if i ever want to do benchmarks i'll use hard numbers
and that still doesn't explain the 150 megs (happy now?) of memory use by firefox. that's RSS, not VSZ
enough of it. until you (or someone else) can give me a solution, or at least a reasonable explanation to that insane memory usage, i stand by what i said.
and quit spouting off your bullshit about caches, and read what i said properly. i already stated this isn't a cache issue, it's a pure memory issue
actually i don't use KDE, and haven't in over two years. i'm currently using openbox, and blackbox before that (moved due to lack of xinerama support)
right now i have a little over 20 applications loaded, of which a lot are terminals. the X server itself takes 10% of my available ram. firefox takes 15% (one window, one tab), openoffice 9% (with one 3 pages document loaded), the java vm 7% (proprietary CVS-like application) and evolution 7% (with lots of mails and accounts, that one is reasonable).
now that's some of the lowest numbers i've ever seen. it isn't rare the X process is around 40-50% along with firefox. that doesn't leave much for the rest of my applications to run, especially since the X process can't be swapped out (always being used).
this box is a p3 500 with 375 megs of ram. it IS insane use of memory.
linux can run on a decade-old computer in shell mode no problem, but don't even think of running X on it (i tried, it "run" but, like running windows xp on a 486, you won't do much with it)
btw distro is debian sid, with the "prelink" package (did the same before i installed it too)
no i dont mean the ram used by the caches. i know about that.
i am talking about (very) simple applications gobbling up 15 megs EACH (took at look at KDE recently?)
firefox on my system routinely use 50% of my total ram (out of 320). that's with only 1 window open and 5-10 tabs. i'd consider that "excessive memory use"
it also does the same when im at work
though linux (the whole deal, whichever distro you use, not just the kernel) in general tend to consume insane amounts of ram, firefox/mozilla is the worst so far
i like my whois command. no fuss, no ugly browser. every info i will ever need on my STDOUT!
Does your Web site launch automatic download prompts?
Microsoft will now suppress downloads not initiated by the user directly
holy shit! does that mean we will get sourceforge/download.com/cnet.com and all the other morons that make you jump through hoops to download a simple file to redeem their ways?
naaaa unlikely
isp here, and we support win95
we also support macos8, and i've been known to help a customer with windows 3.1 (!)
i run a small ISP, and i have the callerid of everyone calling, no matter what their privacy setting says. it even gets logged in my cute little radius database
as someone pointed out, it's a part of the ISDN call setup protocol.
except when your zones get around to morons ISP that force your zones to have a higher TTL than you specified, which unfortunately here is being done by the big player
java for one i know still won't install whitout spending a few hours on it
what about the "too fuckin cumbersome to install a plugin on" line ? because it sure fits firefox nicely
one of the reason for cancelling my cell phone plan and switching to another company was their web site wouldn't work with konqueror/mozilla/firefox.
of course it wasn't the only reason, but it doesn't help them any
that's what you get a PORTABLE space for.
i'm following nanog, and the customer that sued got attributed a portable address space from arin last year. they are just too lazy to renumber
gfs has existed for much longer than google fs. redhat just bought it from sistina and made it GPL (again)
there's also an "OpenGFS" project that was forked from GFS when it was open source. that project seems dead (last update in 2003)
this is a giant step FORWARD. if it can keep my network from being bombarded by all those damned windows viruses it's GOOD no matter what. and i don't even use windows.
i'd say this is the brightest idea microsoft had in the last decade (if they deliver that is)
there are tools to modify the bind config file too, that's not the point. your tools won't help one bit if you want to READ the damn file and understand what it does, unless maybe by converting it to BIND format which defeats the whole purpose (also the bind format just happen to be the dns protocol format)
the documentation might have been there, but it sure as hell wasn't clearly linked on the main page of djbdns, as i remember spending an hour or two looking for it on cr.yp.to
as for the multiple characters. i'm suggesting there is only one character to do something, and use two lines to do the two different things. in the end it's much clearer
i could care less to grep out all the A records. in the 4 years i've been here i never had a use for that. usually im looking for one well defined domain name (like the MX, mail.domain.com) or an IP address (when changing the IP of a server for exemple).
another toy that sound cool on paper but really isn't that useful
i have both a djbdns server (for a customer, 1200 domains or so) and a couple of bind ones (~400 domains).
how the fuck can you say djbdns is easier than bind? if i want an A record in bind it's "IN A" (see, easily understood). if you want the same in djbdns it's some cryptic characters that make no sense (and is, of course, undocumented, or was last time i needed it).
now the best part. there's MULTIPLE characters to do nearly the same thing. if i recall a + is a straight A record, and a @ (i think) is an A+PTR
give me bind anytime, it's MUCH easier. though i'm about to move to powerdns or something with a mysql or ldap backend so customers can edit their zones easily
that's right but their definition of ISP is limited. for them an ISP is an entity that will make sub allocations to other entities.
/48s, by advertising that connectivity through its single aggregated address allocation; and
/48 assignments to other organizations within two years.
i currently work for an "ISP" that is too small to ever need to do sub allocations, but big enough that renumbering would involve about 2 months of work (with appropriate notifications to customers)
here's the appropriate entry in the policy:
5.1.1. Initial allocation criteria
a) be an LIR;
b) not be an end site;
c) plan to provide IPv6 connectivity to organizations to which it will assign
d) have a plan for making at least 200
ipv6 has issues. mostly policy issues but that's pretty major IMHO
/20 (or lower) get to renumber all of their servers if they ever want to switch to another upstream provider, which means you're pretty much fucked if your current ISP goes under or tries to screw you over (yes it happened to me). trust me, renumbering involve more than just setting a new IP address on the server, it's a HUGE pain in the ass. the older your network is, the more of a pain it is
according to ARIN, under ipv6 you can no longer get allocated addresses for your own use. unless you're a big isp (tier 1) that plan to allocate addresses yourself (and lots of them).
that means the small isp that currently has a
i understand the why of that policy (to prevent fragmentation), but it's still annoying