look jackass... he said that your perspective changes on this as you raise a child.
how about trying to refute that? (since you're apparently all about the point by point, but skip the points you cant refute) Get back to us when your oldest child is say... 15.
"your perspective changes" is all he said. he didn't say you're full of shit.
Re:Unfounded Criticism
on
iPods at War
·
· Score: 1
At the end of the day, people don't want soldiers to be "too comfortable" because going out and killing people shouldn't be nice, you should never be comfortable doing it, so you only do it when it's really necessary to. The thought of it being comfortable, rightly or wrongly, makes other people (aka "society") uncomfortable. Not sure whether I agree, but it's certainly understandable.
though snopes debunked it, I still like the story:
A U.S. sniper, asked what he feels when he shoots an al-Qaeda member, replies: "Recoil."
years ago when I ran the news servers at dejanews (yes, I'm old. look at the number of digits in my uid) there was a power event, and the generators kicked in. However the co-lo operators that night were too busy playing warcraft to notice, and the diesel ran dry about 4am.
of course you can do this. Look at datadomain.com.
they expect 20-80x compression because they're marketing themselves as backup to disk (doing repetitive full backups). you get the same patterns over and over again.
and whoever posted the RLE wikipedia article, thank you for understanding the solution.
and no, everything isn't going to compress 25x, but everything will compress some. There are repeated bitstreams in everything. a 64bit string has a finite number of patterns. I don't know how small they chunk it up, but it's beliveable.
the L1400 would more scalable for you if you see growth in the future. The entry cost isn't that much more than an L700 either.
Stay away from the SL500. Had one of those at my last employer. the SL8500 is nice, but has a high entry cost. Scales forever, but very high entry. (have on of those now)
what is your backup window? how long do you want to keep the data around? how much does the data change?
I'd be happy to give you some suggstions, butthose are the questions that need to be answered.
Backup to disk isn't much good if you'reconsidering LTO3. It's faster tham most disk. it only helps with restore latency.
I manage backups for about 300TB of data (mostly oracle, some file server). I'd be happy to through you some suggestions for free, or do a more formal proposal if your management likes that kind of thing. Shoot me an email.
Most restores are the result of hardware or user error. Both of which may end up being emergency restores. No off-hours work there. Also, in a solution as large as this, there will be a dedicated backup network anyway, so you will have 1Gb to work with. That is still a really big limiting factor.
Backup/Restore is what I do. It's my job. I'm happy to share my experiences and architectural opinions on any project.
how fast can you write to storage on any given system? if you put 4 dual port FC cards in a host, and you MAX them all out to the theoretical limit of 256MB/sec (2Gb/sec) then you get 2GB/sec spread across all 8 interfaces. then it doesn't take that long. BUT. while you can add up enough tape drives to get that throughput. getting the backplane to write that fast isn't likely. if you're restoring the data to multiple machines, then you have network issues to deal with.
restoring one PB at 60MB/sec takes about 5000 hours.
let's assume best case. you're allowed to spend money on backups. let's say your solution is going to have 100 servers each with about 10TB on them. and you have 10 LTO3 tape drives. if you get 80MB/sec out of each tape drive (that's about the realistic max of 1Gb ethernet) then it's still 373 hours with all drives spinning at 100%. that's a little over 2 weeks. assuming everything goes right. and depending on how well the data compresses, more than 1000 tapes which hopefully haven't gone bad.
depending on how the data behaves this answer varies greatly. if the data never changes, but more of it is added, then perpetual incrementals is the way to go. if the data changes, then block level incrementals will help you get your backups done, but restores will be a bitch.
The question is really about the restore. once you have a PB on disk, it's going to take weeks to restore. no matter how you back it up. disk to disk is the only option then with a forklift restore.
you have EMC hardware for your storage and you can sleep?
since when should replacing one drive that is marked bad cause 18 others to mark themselves bad. Oh, btw, it turns out, the drive that caused the problems (throwing 18000 FC errors in about 10 minutes) was next to the first one that was replaced (which wasn't really bad) and never got marked as bad.
that's what I call quality.
we don't do hot drive replacements on our clariion's anymore (either of them).
does Xen allow you to change a host OS from having 0.1 of a virtual processor to 2 virtual processors on the fly? I don't know Xen. That's what the dynamic LPAR's do. you can change these things on the fly.
blades don't buy you much. they still price about the same as a bunch of 1U boxes, and if you want hot-swap scsi drives, bladecenter doesn't increase density. if you go with boot from san like you're implying, your density doubles, but I've not known anyone who's had a good experience with boot from san. You can go with the IDE drives and still have the double density, but do you really want ide in your production environment?
if you can't keep the.45 on target, then the 9mm or.40 is the right caliber for you. You're dead on (pun intended) with the need for good follow up shots. Though with the right grip and arm positioning, large caliber pistols are not difficult to keep on target. There are several tactical pistol training academies throughout the states that have the appropriate training available. I'd suggest that anyone who plans on carrying any caliber handgun for personal protection take some tactical pistol training.
You're right about the overpenetration, but you shouldn't be using fmj against an attacker.
The Glock you have is a great pistol. I'm a 1911 fan myself, but the glocks are good too.
carry a.45 and you won't need as much ammo. Roughly half as much as if you're carrying 9mm..40 is a decent comprimise, but tends to go through people. be careful what's behind what you're shooting at.
don't bring a gun if you wish, you may leave that up to the police, but you should keep in mind that the police have no obligation to render aid, or protection. They are not responsable for your safety; you are.
forget the 9mm. you're a civillian. You don't want the guy you're shooting at to fall down so you can handcuff him. You don't want him to get up again. Carry a.45
There have been several good answers for the front end. Here's a good backend architecture.
make sure you virtualize. VMWareESX and Vmotion are very cool. They have tons of info on their site for using virtualization to increase uptime and it's all true. I thought it was a load of BS until I started using it. It's great for DR and multiple sites.
flat backend filestore.... ok, I have nothing to do with these people. I have no stock in the company, I jsut think they have a cool product. http://datadomain.com/
Check out Data Domain if you're going to use flat filesystem for the filestore. They use bitpattern matching to provide pseudo single instance store, and (they claim) 20x compression (though with this technology and something like mail, you could probably approach 8x easially.
Their products do remote replication so you can have your multpile sites with the same mailstores.
also figure storage.... 1M users, figure average message size is 15K (in a single instance store system, no SIS, figure 75K). Figure everyone is going to have 1000 messages in their inbox.
so that's 15-75TB if you could limit mailbox size reasonably, you could probably get away with the DataDomain DD460 without too much hassle. put one at each site, set up asynchronous replication, buy 2 extras for backups in different locations and offset their synchro schedules. if you need a message delete more than a few weeks ago, tell whoever wants it to go ask the corporate lawyers why you shouldn't keep email on tapes. If they really want to back it up to tape, email me and I'll build you an architecture on paper.
if you don't use a DD product for your backend, look at pillardata.com you could build a 20TB system for about $6/GB and when you fill it up, expand it for about $3.50/GB (in 4TB chunks)
I do storage management for a living. I have about 160TB of accessible storage spinning right now. Backend I can do off the top of my head. Front end is best left to others, but the backend is always the same.
Slashdot Mirror
look jackass... he said that your perspective changes on this as you raise a child.
how about trying to refute that? (since you're apparently all about the point by point, but skip the points you cant refute) Get back to us when your oldest child is say... 15.
"your perspective changes" is all he said. he didn't say you're full of shit.
At the end of the day, people don't want soldiers to be "too comfortable" because going out and killing people shouldn't be nice, you should never be comfortable doing it, so you only do it when it's really necessary to. The thought of it being comfortable, rightly or wrongly, makes other people (aka "society") uncomfortable. Not sure whether I agree, but it's certainly understandable.
though snopes debunked it, I still like the story:
A U.S. sniper, asked what he feels when he shoots an al-Qaeda member, replies: "Recoil."
find a local microbrew.
sometimes they do kick in, but it doesn't matter.
years ago when I ran the news servers at dejanews (yes, I'm old. look at the number of digits in my uid) there was a power event, and the generators kicked in. However the co-lo operators that night were too busy playing warcraft to notice, and the diesel ran dry about 4am.
fecking whiney users. create a complex password that isn't hard to remember. use keyboard patterns, or just memorize.
/.;lpo)(
examples of keyboard patterns that are hard to guess, easy to remember
a1s2D#F$
zqxwCEVR
you get the idea.
it's the beginning of the starship troopers style of MI.
RTFA.
of course you can do this. Look at datadomain.com.
they expect 20-80x compression because they're marketing themselves as backup to disk (doing repetitive full backups). you get the same patterns over and over again.
and whoever posted the RLE wikipedia article, thank you for understanding the solution.
and no, everything isn't going to compress 25x, but everything will compress some. There are repeated bitstreams in everything. a 64bit string has a finite number of patterns. I don't know how small they chunk it up, but it's beliveable.
the L1400 would more scalable for you if you see growth in the future. The entry cost isn't that much more than an L700 either.
Stay away from the SL500. Had one of those at my last employer. the SL8500 is nice, but has a high entry cost. Scales forever, but very high entry. (have on of those now)
what is your backup window?
how long do you want to keep the data around?
how much does the data change?
I'd be happy to give you some suggstions, butthose are the questions that need to be answered.
Backup to disk isn't much good if you'reconsidering LTO3. It's faster tham most disk. it only helps with restore latency.
I manage backups for about 300TB of data (mostly oracle, some file server). I'd be happy to through you some suggestions for free, or do a more formal proposal if your management likes that kind of thing. Shoot me an email.
just goes to show that most humans don't really understand the meaning of the words they use.
I've thought for many years that the use of fold to mean 'multiplied by' was moronic. I understand it's mis-use, but everytime I see it it's annoying.
For the most part though SMTP is easier to track because most servers do store the senders IP
the majority of usenet servers also stick the sender's IP addr in the headers.
Most restores are the result of hardware or user error. Both of which may end up being emergency restores. No off-hours work there. Also, in a solution as large as this, there will be a dedicated backup network anyway, so you will have 1Gb to work with. That is still a really big limiting factor.
Backup/Restore is what I do. It's my job. I'm happy to share my experiences and architectural opinions on any project.
how fast can you write to storage on any given system? if you put 4 dual port FC cards in a host, and you MAX them all out to the theoretical limit of 256MB/sec (2Gb/sec) then you get 2GB/sec spread across all 8 interfaces. then it doesn't take that long. BUT. while you can add up enough tape drives to get that throughput. getting the backplane to write that fast isn't likely. if you're restoring the data to multiple machines, then you have network issues to deal with.
restoring one PB at 60MB/sec takes about 5000 hours.
let's assume best case. you're allowed to spend money on backups. let's say your solution is going to have 100 servers each with about 10TB on them. and you have 10 LTO3 tape drives. if you get 80MB/sec out of each tape drive (that's about the realistic max of 1Gb ethernet) then it's still 373 hours with all drives spinning at 100%. that's a little over 2 weeks. assuming everything goes right. and depending on how well the data compresses, more than 1000 tapes which hopefully haven't gone bad.
how long can you be without the data....
depending on how the data behaves this answer varies greatly. if the data never changes, but more of it is added, then perpetual incrementals is the way to go. if the data changes, then block level incrementals will help you get your backups done, but restores will be a bitch.
The question is really about the restore. once you have a PB on disk, it's going to take weeks to restore. no matter how you back it up. disk to disk is the only option then with a forklift restore.
you have EMC hardware for your storage and you can sleep?
since when should replacing one drive that is marked bad cause 18 others to mark themselves bad. Oh, btw, it turns out, the drive that caused the problems (throwing 18000 FC errors in about 10 minutes) was next to the first one that was replaced (which wasn't really bad) and never got marked as bad.
that's what I call quality.
we don't do hot drive replacements on our clariion's anymore (either of them).
it's not even the cheapest. check out nexsan atabeast
does Xen allow you to change a host OS from having 0.1 of a virtual processor to 2 virtual processors on the fly? I don't know Xen. That's what the dynamic LPAR's do. you can change these things on the fly.
blades don't buy you much. they still price about the same as a bunch of 1U boxes, and if you want hot-swap scsi drives, bladecenter doesn't increase density. if you go with boot from san like you're implying, your density doubles, but I've not known anyone who's had a good experience with boot from san. You can go with the IDE drives and still have the double density, but do you really want ide in your production environment?
if you can't keep the .45 on target, then the 9mm or .40 is the right caliber for you. You're dead on (pun intended) with the need for good follow up shots. Though with the right grip and arm positioning, large caliber pistols are not difficult to keep on target. There are several tactical pistol training academies throughout the states that have the appropriate training available. I'd suggest that anyone who plans on carrying any caliber handgun for personal protection take some tactical pistol training.
You're right about the overpenetration, but you shouldn't be using fmj against an attacker.
The Glock you have is a great pistol. I'm a 1911 fan myself, but the glocks are good too.
carry a .45 and you won't need as much ammo. Roughly half as much as if you're carrying 9mm. .40 is a decent comprimise, but tends to go through people. be careful what's behind what you're shooting at.
don't bring a gun if you wish, you may leave that up to the police, but you should keep in mind that the police have no obligation to render aid, or protection. They are not responsable for your safety; you are.
c rime.com/policeprotection.html
http://flyservers.registerfly.com/members5/police
forget the 9mm. you're a civillian. You don't want the guy you're shooting at to fall down so you can handcuff him. You don't want him to get up again. Carry a .45
There have been several good answers for the front end. Here's a good backend architecture.
make sure you virtualize. VMWareESX and Vmotion are very cool. They have tons of info on their site for using virtualization to increase uptime and it's all true. I thought it was a load of BS until I started using it. It's great for DR and multiple sites.
flat backend filestore....
ok, I have nothing to do with these people. I have no stock in the company, I jsut think they have a cool product. http://datadomain.com/
Check out Data Domain if you're going to use flat filesystem for the filestore. They use bitpattern matching to provide pseudo single instance store, and (they claim) 20x compression (though with this technology and something like mail, you could probably approach 8x easially.
Their products do remote replication so you can have your multpile sites with the same mailstores.
also figure storage.... 1M users, figure average message size is 15K (in a single instance store system, no SIS, figure 75K). Figure everyone is going to have 1000 messages in their inbox.
so that's 15-75TB if you could limit mailbox size reasonably, you could probably get away with the DataDomain DD460 without too much hassle. put one at each site, set up asynchronous replication, buy 2 extras for backups in different locations and offset their synchro schedules. if you need a message delete more than a few weeks ago, tell whoever wants it to go ask the corporate lawyers why you shouldn't keep email on tapes. If they really want to back it up to tape, email me and I'll build you an architecture on paper.
if you don't use a DD product for your backend, look at pillardata.com you could build a 20TB system for about $6/GB and when you fill it up, expand it for about $3.50/GB (in 4TB chunks)
I do storage management for a living. I have about 160TB of accessible storage spinning right now. Backend I can do off the top of my head. Front end is best left to others, but the backend is always the same.
beat me to it!