And I agree. However I have spoken to way too many people who not only did not have redundant backups, but were using the same media well past expected life. Somehow it became MY fault when they could not restore.
As well, even if you do have offsite backups, that does increase the time to restore. Backups are not just for disaster recovery. They can be (and in my company are) used for file control and restoration.
Voodoo2 cards were not video cards. They were only 3D accelerators. They had to tie into a graphics card through a pass-through cable. The closest you could come to that (which was slower in 3D) was the Voodoo Banshee.
Voodoo3 cards were graphics cards AND accelerators.
He was talking about how more density would end up meaning that you would use fewer tapes. Thus, it implies that a single tape failure would expose a greater amount of data to loss. It is very simple. If your backup requires that you use 10 tapes and you lose 1 tape, you lose 1/10th of your data. If your backup used 1 tape and you lose 1 tape. You have then lost ALL of your data.
I think he is talking about the fragility of the medium. Tape is thin, and easily damaged inside the read/write heads of a drive. If you ever had a drive go south reading from a B/U tape during a critical restore, you know what I mean. The issue is hopefully it will be faster, cheaper etc so you can copy those tapes before disaster hits twice.
If you read the article, it sounded like some sort of apologistic approach anyhow:
"If star A is in Uranus while star B is in Neptune, and you look at a pr0n site, then there is a chance that the girl's boobies will appear larger than they are"
Basically it states that if the thermals are bad, and you cause a hotspot on the CPU, that there MAY be corruption in the data processed.
Thats a bit of a stretch. Its not good, but its not major.
That proves nothing. It just means that there is a lot of penetration in symantec products. I am not a fan of a lot of their products, but Brightmail is not on the list of crap they have put out. They BOUGHT brightmail and made it their own.
Just an FYI, looking up "Linux Error" in google gets me 72,800,000 hits. Looking up "Sexual Error" gets me 15,600,000 hits and "google error" gets me 65,800,000 hits.
at work. However, have you considered instead of using brightmail on the exchange server, only use the foldering agent and set up brightmail filters as your MX record (top level) and have them relay the mail to your exchange? We have about >95% catch rate. You can set them up running on Windows with IIS SMTP, Linux with sendmail or Solaris with sendmail. As cheap as brightmail is and as good as it has worked for my company, I would keep it. My suggestion would be to use the "Suspected Spam" option and set the threshold to 62.
The one thing I would suggest is if it is a windows based gateway filter, as described above, reboot it weekly (works really well if you can afford 2 boxes, since BM doesn't charge by server, CPUs or anything, but rather how many clients you have it filter for) or at least schedule scripts to restart tomcat (net stop tomcat... net start tomcat...)
If on a *Nix box, just cron tomcat restarts.
You live in a fantasy land. Your kids will do what they will do unless you tie them to a bed or chain them to a radiator. You only THINK they will not do these things. They will do them anyway; you will be proven wrong.
Yes, you are correct. Think about how it would streamline development/testing. You want to try out a new feature or patch to an application. You bring a snap of your VM to your testing server. You launch your VM of the production server, make the change, and validate. When done, you snap that over to the real box and you are live. You don't have to implement the change twice. If it is trivial, not a big deal, but if it is a major issue like a multi hour recompile, this could save tons of time.
Very interesting points. My company is moving towards virtual servers (not MS Virtual Server) for some of those very reasons. Right now, if a server fails we need to either reinstall the OS and restore from backup tape or replace the failed hardware.
In the new senario, any one server failure will simply result in the VMs being down for moments (we're not paying for the expensive up all the time jazz, hard enough to get the money for this project).
When it is time to upgrade, we simply roll in faster hardware and move the VMs over to it. If we want to test upgrades to the guest OSes, we can simply take a snapshot of the LUN it is on, test, and if we like it, fully populate it to the SAN. If we don't, we roll it back and no harm done.
You are thinking small. I am not saying "Aquire Novell for Novell's current assets and asset structure. I am saying use Novell's assets along with your own to become greater than the sum of the parts. You could sell the pie and the pie plate with the fork too. SuSE could easily become the defacto standard for Oracle Databases. OpenExchange with Oracle DB back ends would be sales of more databases, etc; a real viable chance to take on MS in that arena.
Perhaps if Oracle bought Novell they would then have the power to give MS a true run for its money. Novell's Desktop Linux, OpenExchange with an Oracle DB back end and Evolution as the email client. OpenOffice/etc for the office suite. Opera or Firefox as their browser and Oracle, while still supporting Unices and Windows, has their own tuned Linux distro (SLES) to push their databases out on. This would be good competition for both Redhat AND Microsoft.
considering I am the only one who logs in as root and I have done nothing to it besides....su to root and on occasion launch X, I should have no reason to think that the profile (including the X session files) become corrupt. This is all on an IBM 1U box that otherwise runs fine. Luckily my regular user and root's profile are strikingly similar. I can just rewrite the files that become corrupted faster than going to a backup tape. After the third time, I kept a backup of the files tar'd up in a safe place just in case.
Well, can't say about this for everyone, but at my company (predominantly windows for the back end) we are retiring our single Redhat server (ES 3.0) and putting up 3 Novell/SuSE (SLES 9) servers. I found SuSe to be superior to Redhat in many ways. For one, RH ES 3.0 had a nasty habit of having root's profile become corrupted. No other user had this issue. While the machine was very stable otherwise, I found SuSE better in many many ways. The part about not being under the gun for updates was a big sell for us.
Someone never heard of a program called AVP (now Kaspersky Antivirus) that would actually remove NYB and repair the boot sector. The alternative would be to run Fdisk/mbr which would remove the infection. This was NOT a good idea with a "Data diddler" infection like ripper, but NYB would be removed in this fashion.
This isn't news. It is just going to another realm. This type of logical gerrymandering happens in other contexts all the time. In efforts to put constraints (read: ban) guns, they show how many "Children" are killed each year in gun related deaths. They include "Children" up to 20 or 21yrs old in many of these research examples.
They also do not reject samples of "Child" gang members who are shooting at each other, etc. I guess we need locks on all our computers so that kids cannot accidentally kill themselves by seeing a boobie.
because there is no free lunch. In many cases, context matters as much as coding ability or technology ability. If there is a language, cultural or contextual problem, the project will bleed money in other areas
in many cases, HR won't let you come in the door unless you have certifications. I have seen jobs for Linux Admins where they expect you to have MCSE and knowledge of Windows 2000 Adv server where the job role says NOTHING about windows besides that little part.
The way that contracts work, if Party A drafts a contract to enter into agreement with Party B, should there be any ambiguity the contract is interpreted to the benefit of Party B as they did not draft the contract. This protects parties from being take advantage of by ambiguous statements made on purpose. If there is room for interpretation, then the contract could benefit Party A or Party B. Since Party A drafted the contract, it would be unfair for the ambiguity to benefit them.
Slashdot Mirror
Amen,
If you want to "hide" behind the First......be prepared to use the Second. That is why it was put there.
Those who would hammer their guns into plowshares will plow for those that did not.
And I agree. However I have spoken to way too many people who not only did not have redundant backups, but were using the same media well past expected life. Somehow it became MY fault when they could not restore. As well, even if you do have offsite backups, that does increase the time to restore. Backups are not just for disaster recovery. They can be (and in my company are) used for file control and restoration.
Voodoo2 cards were not video cards. They were only 3D accelerators. They had to tie into a graphics card through a pass-through cable. The closest you could come to that (which was slower in 3D) was the Voodoo Banshee. Voodoo3 cards were graphics cards AND accelerators.
He was talking about how more density would end up meaning that you would use fewer tapes. Thus, it implies that a single tape failure would expose a greater amount of data to loss. It is very simple. If your backup requires that you use 10 tapes and you lose 1 tape, you lose 1/10th of your data. If your backup used 1 tape and you lose 1 tape. You have then lost ALL of your data.
I think he is talking about the fragility of the medium. Tape is thin, and easily damaged inside the read/write heads of a drive. If you ever had a drive go south reading from a B/U tape during a critical restore, you know what I mean. The issue is hopefully it will be faster, cheaper etc so you can copy those tapes before disaster hits twice.
Same thing of device files on HP-UX. I wonder when those two will merge, as per the Open Letter to Mr. Hurd. (no relation to the kernel).
what would you say to someone who says you:
/dev/null
mv $file
??
If you read the article, it sounded like some sort of apologistic approach anyhow: "If star A is in Uranus while star B is in Neptune, and you look at a pr0n site, then there is a chance that the girl's boobies will appear larger than they are" Basically it states that if the thermals are bad, and you cause a hotspot on the CPU, that there MAY be corruption in the data processed. Thats a bit of a stretch. Its not good, but its not major.
Just an FYI, looking up "Linux Error" in google gets me 72,800,000 hits. Looking up "Sexual Error" gets me 15,600,000 hits and "google error" gets me 65,800,000 hits.
at work. However, have you considered instead of using brightmail on the exchange server, only use the foldering agent and set up brightmail filters as your MX record (top level) and have them relay the mail to your exchange? We have about >95% catch rate. You can set them up running on Windows with IIS SMTP, Linux with sendmail or Solaris with sendmail. As cheap as brightmail is and as good as it has worked for my company, I would keep it. My suggestion would be to use the "Suspected Spam" option and set the threshold to 62. The one thing I would suggest is if it is a windows based gateway filter, as described above, reboot it weekly (works really well if you can afford 2 boxes, since BM doesn't charge by server, CPUs or anything, but rather how many clients you have it filter for) or at least schedule scripts to restart tomcat (net stop tomcat... net start tomcat...) If on a *Nix box, just cron tomcat restarts.
Mod Parent UP
You live in a fantasy land. Your kids will do what they will do unless you tie them to a bed or chain them to a radiator. You only THINK they will not do these things. They will do them anyway; you will be proven wrong.
Yes, you are correct. Think about how it would streamline development/testing. You want to try out a new feature or patch to an application. You bring a snap of your VM to your testing server. You launch your VM of the production server, make the change, and validate. When done, you snap that over to the real box and you are live. You don't have to implement the change twice. If it is trivial, not a big deal, but if it is a major issue like a multi hour recompile, this could save tons of time.
Very interesting points. My company is moving towards virtual servers (not MS Virtual Server) for some of those very reasons. Right now, if a server fails we need to either reinstall the OS and restore from backup tape or replace the failed hardware.
In the new senario, any one server failure will simply result in the VMs being down for moments (we're not paying for the expensive up all the time jazz, hard enough to get the money for this project).
When it is time to upgrade, we simply roll in faster hardware and move the VMs over to it. If we want to test upgrades to the guest OSes, we can simply take a snapshot of the LUN it is on, test, and if we like it, fully populate it to the SAN. If we don't, we roll it back and no harm done.
You are thinking small. I am not saying "Aquire Novell for Novell's current assets and asset structure. I am saying use Novell's assets along with your own to become greater than the sum of the parts. You could sell the pie and the pie plate with the fork too. SuSE could easily become the defacto standard for Oracle Databases. OpenExchange with Oracle DB back ends would be sales of more databases, etc; a real viable chance to take on MS in that arena.
Perhaps if Oracle bought Novell they would then have the power to give MS a true run for its money. Novell's Desktop Linux, OpenExchange with an Oracle DB back end and Evolution as the email client. OpenOffice/etc for the office suite. Opera or Firefox as their browser and Oracle, while still supporting Unices and Windows, has their own tuned Linux distro (SLES) to push their databases out on. This would be good competition for both Redhat AND Microsoft.
are you including CALs in your pricing of both?
considering I am the only one who logs in as root and I have done nothing to it besides....su to root and on occasion launch X, I should have no reason to think that the profile (including the X session files) become corrupt. This is all on an IBM 1U box that otherwise runs fine. Luckily my regular user and root's profile are strikingly similar. I can just rewrite the files that become corrupted faster than going to a backup tape. After the third time, I kept a backup of the files tar'd up in a safe place just in case.
Well, can't say about this for everyone, but at my company (predominantly windows for the back end) we are retiring our single Redhat server (ES 3.0) and putting up 3 Novell/SuSE (SLES 9) servers. I found SuSe to be superior to Redhat in many ways. For one, RH ES 3.0 had a nasty habit of having root's profile become corrupted. No other user had this issue. While the machine was very stable otherwise, I found SuSE better in many many ways. The part about not being under the gun for updates was a big sell for us.
Someone never heard of a program called AVP (now Kaspersky Antivirus) that would actually remove NYB and repair the boot sector. The alternative would be to run Fdisk /mbr which would remove the infection. This was NOT a good idea with a "Data diddler" infection like ripper, but NYB would be removed in this fashion.
This isn't news. It is just going to another realm. This type of logical gerrymandering happens in other contexts all the time. In efforts to put constraints (read: ban) guns, they show how many "Children" are killed each year in gun related deaths. They include "Children" up to 20 or 21yrs old in many of these research examples.
They also do not reject samples of "Child" gang members who are shooting at each other, etc. I guess we need locks on all our computers so that kids cannot accidentally kill themselves by seeing a boobie.
why not just rm *
because there is no free lunch. In many cases, context matters as much as coding ability or technology ability. If there is a language, cultural or contextual problem, the project will bleed money in other areas
in many cases, HR won't let you come in the door unless you have certifications. I have seen jobs for Linux Admins where they expect you to have MCSE and knowledge of Windows 2000 Adv server where the job role says NOTHING about windows besides that little part.
The way that contracts work, if Party A drafts a contract to enter into agreement with Party B, should there be any ambiguity the contract is interpreted to the benefit of Party B as they did not draft the contract. This protects parties from being take advantage of by ambiguous statements made on purpose. If there is room for interpretation, then the contract could benefit Party A or Party B. Since Party A drafted the contract, it would be unfair for the ambiguity to benefit them.
But IANAL