In the Win3.x and Win9x and NT4 days I built and maintained computers labs at a community college.
For Win3.x, WP5.x and 6.x shipped with a msvrt.dll and one other MS provided DLL that was identical binary (and version number) to what was shipped with Windows itself. The WordPerfect installer replaced the Windows installed copy with it's own copy where the timestamp was the only difference.
As soon as you installed MS-Office these two DLLs (in the windows directory) were replaced with different binaries which reported the same version number.
Once installed, WordPerfect would be slower and more crash prone.
I don't know if this was a planned/malicious change on the part of Microsoft or just crappy version control development, you decide.
They should have done brand analysis (without naming the brand) and also rpm analysis.
From the article..
3.2 Manufacturers, Models, and Vintages Failure rates are known to be highly correlated with drive models, manufacturers and vintages [18]. Our results do not contradict this fact. For example, Figure 2 changes significantly when we normalize failure rates per each drive model. Most age-related results are impacted by drive vintages. However, in this paper, we do not show a breakdown of drives per manufacturer, model, or vintage due to the proprietary nature of these data.
reliability of an internet service is not always dependent on bandwidth used.
Most times it depends on your ISP and their upstream ISPs' network, quality of service, etc.
Note: it is possible you are reaching your real bandwidth limit (what the ISP makes available to you) at different times which causes delays resulting in timeouts. Might be volume of traffic over the VPN link or it might be non-VPN traffic to the rest of the internet.
If you want more bandwidth (hedged bet for the future) and better reliability, then what you want is redundant ISPs and VPN concentrators that support redundant ISPs and more importantly support load sharing.
I have no relationship with the following company except I evaluated one of their products earlier this year for non-VPN redundant ISP problem.. http://www.astrocorp.com/
Given what you say about large price delta between DSL@512 w/1:2 vs 1Mbps w/1:4, you would be better off with two DSLs each @512 w/1:2 from two different ISPs. Note: The key to redundancy is making sure each of the ISPs you choose don't share the same upstream ISPs. Not sure of the market in India but this is hard in many locations where the local telco is the only upstream provider.
The Backup server or cluster of servers store 20KB blocks keyed to the block's SHA-1 hash.
Smart agents on each backup client chunks each new file to be backed up into 20KB blocks and calculates SHA-1 hashes which it compares against the backup server.
If the block is new (not on the backup server) the block itself is transfered. If the block is old, the backup server stores an extra reference to the block for the client/file.
The end result is.. a) a 1000 windows backup clients will result in only one copy of a windows dll being saved b) every full backup is like an differential in terms of size/speed. c) you have weeks (if not months) of daily backups of all backup clients stored in 110% of total backup size. d) the backup agent on the client has a larger footprint and requires more CPU while running.
Put the backup servers at a remote site with a high speed link and you have disaster recovery as well. If the high speed link isn't an option, there is support for remote replication; requires another backup server (or cluster) of same size.
This is the way all backups will be done in the future.
BUT IF you want weekly, monthly, quarterly, yearly backups then a virtual tape library (VTL) is a better option. For most servers, the change in the dataset is small and gradual so a VTL stores one full compressed back + diffs for incremental/differential/full backups. Also, VTLs look for redundant data across servers; 10 similar linux servers will have the almost identical binaries.
I am currently looking at http://www.datadomain.com/ VTL to replace a 72 slot dual drive LTO 1st gen library.
A VTL costs a bit less than a regular tape library + all the tapes you need but the increased throughput and no more tape handling is what makes it worth it.
I have yet to see benchmarks comparing Linux2.6, Solaris10 with BSD variants. Anyone have any links?
Most Oracle installations aren't run on Solaris.. a) Granted larger Oracle installations (8Processors or more) are on SPARC/Solaris. b) Many small to medium sized installations are run on x86/Linux. Has been this way for a few years now, ever since Oracle started supporting Linux really aggressively.
Solaris's major advantage is standardized kernel, kernel APIs and system libraries. It allows application developers to better target the platform they want to develop on and support and for how long. In the commercial space this is a big advantage for Solaris.
Where Solaris fails compared to say Redhat (note I am talking about commercial version) is how easy it is to manage the system. Want to apply the latest patches that have been approved by Redhat? up2date For Solaris? go to sun's site hunt for the right page that will list the latest patch cluster. verify this patch cluster doesn't break any of your Sun applications (e.g. SunOne messaging 5.2 sp2 has problems with the latest patch cluster for solaris 9).
Why should a user have to hunt for this information? Why should I have to phone support and have them hunt for it? Why isn't this information on the patch cluster's download page? Why doesn't Solaris have a patch management system that covers all Sun products installed on a server?
US Gas consumption has steadly gone up for the past 10 yrs.
The reason prices are rising is increased demand from other nations, e.g. China (you know the country that makes all those cheap goods you can buy at walmart et al).
Which brand on motherboard was this nforce2 board?
YMMV but cheap manufacturer's cut corners on board design, caps, etc using the same chipset, just sell a motherboard that is $50 cheaper than the rest.
Bottom line: I was not getting my money's worth. Not by a long shot. And it got far worse when I supported a wife and kids. Despite keeping three other individuals off the public dole, my tax rate remained about the same, with extremely small additional non-refundable credits. I've done the analysis in another post.
I don't know what you are doing wrong, my marginal tax rate drops (every year for at least the last 6 years) from 39% to ~23% when I add my family to the dependents section + RRSPs (401K in US).
I would agree with you that taxes paid are easily more than triple the cost of benefits received. Part of this is the inheirent cost of goverment but the big chunk is most canadians are less well off than you, you subsidize their healthcare and education, etc.
Your parent's may have had a hard life but without socialism you probably wouldn't have been able to go to university. My mother never went to work when I was growing up (70s and 80s) and money was tight through all my childhood. I know my siblings and I wouldn't be able to go to uni. (3 of us all started in a 2yr timeframe) if it were for.. a) low tuition b) gov. grants c) but mostly gov. student loans (interest free until 6months after graduation).
Specific to university/college (higher education)..
I used to think like you did, until I became aware of the budgeting process at my college. Currently, students pay less than 25% the actual cost of the education they receive. In the 80s (gov. transfer payments were much higher) the percentage was less than 10%.
The remaining is/was covered by gov. funds in one form or another (all derived from taxes).
By the way, socialism took hold earlier than Trudeau.
One could argue Diefenbaker started it at a federal level. http://canadaonline.about.com/cs/primemini ster/p/p mdiefenbaker.htm
Initiated earlier by Tommy Douglas at a provincial level in Saskatewan.. http://www.weyburnreview.com/tommydo uglas/tommydou glasintro.html
Your version of history isn't entirely correct/complete..
1. Yes the raja (roughly translates to king but is more like a local lord) of kashmir wanted to maintain his monarchy (didn't want either pakistan or india as both's political systems were shaping up to be purely parliamentary, i.e. no monarchy).
2. Under the terms of partition (agreed upon by muslim and hindu political parties.. a) in states where a popular majority (muslim or hindu) exists, that decides which country that state joins (geographical limitations allowing). b) in a state where no clear majority exists, the ruler of the state will decide.
3. In the case of Kashmir.. a) there was and still is a muslim majority. b) when the political reality of keeping kashmir as an independent monarchy wasn't possible, the raja choose india (against partition rules).
That is one of the major seeds of discontent that has lead to two wars between the two countries. All this is history.
For the last 20 years, in an attempt to stamp out terrorism as an excuse, the indian army has been targeting any muslims who raise a political voice; arrest, torture without plausible cause (sometimes leading to dead). It has only created a new generation of terrorists (freedom fighters) and only leads to lend weight to the calls for seperation from india.
I can only dig up one reference right now: http://www.amnesty.org/results/is/eng?queryT ype=0& searchIn=0&query=kashmir&start=1&num=10&max=25&sor tBy=date
university scholarship: You think your 1980's university fees covered the cost of the education you got? Goverment transfer payments covered over 95% of the costs (of the university) for you. Don't you think you have an obligation to pay for the next generation? Just like the previous generation did?
You're seriously comparing an
abandoned webmail program to an entire OS?
You sir are trolling.
I know of two colleges in the lower mainland of BC, Canada that are using NetMail in production for the last 4years (provide email to all students). One of these colleges just bought a new portal system that comes bundled with SunOne messaging server (email integrated portal) and they still stuck with NetMail for email. Why? Out of the box it is designed with features that make an admin's life easier (think seamless email quotas, etc).
w.r.t to the rest of your comments about Sun.. They still stick to an attitude and culture that is elitiest and down right snotty. Reminds me of IBM in the 70s. I don't know if Sun will survive; they claim they want to be a services company but still want to sell hardware+software bundles while VARs provide the real services.
In theory, you're running a system where "crashes" doesn't always imply "hard locks". In that case you just ssh in and reboot.
Specifically thinking about when the HW intermittently hard locks, have seen it happen on a server that I build from desktop class parts. Heck I have seen it happen on a G4 XServe (turned out to be bad ram; for 4wks server would cause the clients to crash intermittently).
Better yet, you either fix it or junk it - even worse is a computer that subtly corrupts data.
Yes best to junk it but what if the cause is a defect in the drivers or worse in the hardware? Then any same model spare is useless too:(
At work we used to buy OEM SuperMicro servers from a local reseller for PC class servers.
Unix was mostly AIX with a bit of Sun Sparc with support provided by IBM or a local Sun reseller.
For PC class servers we initially used Mylex RAID controllers only. They were great boards and great RAID controllers. But a few problems.. a) supermicro server's have a very select list of compatible memory (vendor/model/batch). This added significantly to the cost. b) Mylex RAID controllers can be temperamental to SCSI timing issues and their driver support does NOT expose RAID events (disk failure) to the host OS. c) Mylex driver support for OSes other than Windows wasn't the greatest, driver's took many months to come out and usually required a firmware upgrade of the RAID array at the same time. This is fine on a new server that doesn't have any data but it is a big deal when you have a in production server with loads of data. If something screws up in the firmware update you can loose the RAID array, have to restore your data from backup which for large servers means at least a day of down time.
Also, at the time we were buying one class of server for all our server needs.
So we decided to switch to two class of PC servers.. 1. Large: Dual/Quad Xeon SuperMicro Servers with Promise Ultratrak RM8000 external RAID arrays. 2. Small: Single P4 + Kingston ECC RAM + Promise IDE software RAID (RAID1) from a local whitebox reseller.
This worked well for a while except.. 1. Large Servers: a) Promise Ultratrak would hard lock when installing Redhat Enterprise 2.1. Something to do with Redhat tweaking the SCSI timings of the onboard SCSI controller (connects to external RAID) to suppport clustering. This was fixed by Promise with a firmware update about 14months after we encountered this problem. b) WesternDigital's Accoustic Management would cause the Promise to timeout a drive causing the array to go critical, causing a rebuild to the hotspare, slowing the array down (50%) for 10+hrs (RAID 1+0 of eight 200GB disks). This was fixed about 8months later with a fix from western Digital. c) The PRomise RAID array would not properly handle failed cached writes to disk, causing either corruption or drive to be mark bad (instead of marking the block bad and remapping it). Without disk caching enabled the performance of the Promise RAID array sucks.
2. Small: a) Promise IDE RAID card (software RAID) was not supported well on most OSes except Windows. Their Linux driver left a lot to be desired. b) One time we installed an Adaptec 2940UW scsi card (something we had lying around) to attach a local tape drive. Initially it worked great for a few months then the card started to die; the server would intermittenly reboot without any cause. We lived with this for a few months (all the time riding the local reseller's ass) until finally the card die gloriously, taking the motherboard and power supply with it. I have to give props to the reseller (Cybex Systems Inc. in Vancouver, BC, Canada), they were more than understanding and replaced everything in ASAP (sameday).
After all these problems it was decided we needed a top-tier name brand for our servers. After an RFP we decided on Dell; they were the cheapest by 20%, a local parts depot (this is important), their 4hr warranty was target for completed repair not just response (what HP/IBM have) and most importantly other companies provided a favourable recommendation.
Has it been great?
Well not exactly..
Two of the first few servers we purchased were for a two node Linux webserver setup where replication software would keep the two server's filesystems insync. On the Dell PowerEdge 2650, the embedded RAID controller is an Adaptec (some variation of single channel Ultra160 family). Unfortunately, adaptec controller's have many hardware defects that the driver has to compenstate for (most hard is like this, some more than others). With the replication software installed, within a
Slashdot Mirror
In the Win3.x and Win9x and NT4 days I built and maintained computers labs at a community college.
For Win3.x, WP5.x and 6.x shipped with a msvrt.dll and one other MS provided DLL that was identical binary (and version number) to what was shipped with Windows itself. The WordPerfect installer replaced the Windows installed copy with it's own copy where the timestamp was the only difference.
As soon as you installed MS-Office these two DLLs (in the windows directory) were replaced with different binaries which reported the same version number.
Once installed, WordPerfect would be slower and more crash prone.
I don't know if this was a planned/malicious change on the part of Microsoft or just crappy version control development, you decide.
Great choice of words...
"Israel colonises the place"
Sort of like how the white man colonized north america, right?
They should have done brand analysis (without naming the brand) and also rpm analysis.
From the article..
How do you deal with new exe or dll installed as part of Windows update, anti-virus update, etc?
e.g. IE7 will soon be released via Windows Update.
reliability of an internet service is not always dependent on bandwidth used.
Most times it depends on your ISP and their upstream ISPs' network, quality of service, etc.
Note: it is possible you are reaching your real bandwidth limit (what the ISP makes available to you) at different times which causes delays resulting in timeouts. Might be volume of traffic over the VPN link or it might be non-VPN traffic to the rest of the internet.
If you want more bandwidth (hedged bet for the future) and better reliability, then what you want is redundant ISPs and VPN concentrators that support redundant ISPs and more importantly support load sharing.
I have no relationship with the following company except I evaluated one of their products earlier this year for non-VPN redundant ISP problem..
http://www.astrocorp.com/
Given what you say about large price delta between DSL@512 w/1:2 vs 1Mbps w/1:4, you would be better off with two DSLs each @512 w/1:2 from two different ISPs.
Note: The key to redundancy is making sure each of the ISPs you choose don't share the same upstream ISPs.
Not sure of the market in India but this is hard in many locations where the local telco is the only upstream provider.
is http://www.avamar.com/
The Backup server or cluster of servers store 20KB blocks keyed to the block's SHA-1 hash.
Smart agents on each backup client chunks each new file to be backed up into 20KB blocks and calculates SHA-1 hashes which it compares against the backup server.
If the block is new (not on the backup server) the block itself is transfered.
If the block is old, the backup server stores an extra reference to the block for the client/file.
The end result is..
a) a 1000 windows backup clients will result in only one copy of a windows dll being saved
b) every full backup is like an differential in terms of size/speed.
c) you have weeks (if not months) of daily backups of all backup clients stored in 110% of total backup size.
d) the backup agent on the client has a larger footprint and requires more CPU while running.
Put the backup servers at a remote site with a high speed link and you have disaster recovery as well.
If the high speed link isn't an option, there is support for remote replication; requires another backup server (or cluster) of same size.
This is the way all backups will be done in the future.
Most VTL have support for remote replication but it is usually extra.
Depending on requirements, broadband link is all that is needed.
More bandwidth can be had with point to point microwave links (300km limit).
IF you want long term archiving, still need tape.
BUT IF you want weekly, monthly, quarterly, yearly backups then a virtual tape library (VTL) is a better option. For most servers, the change in the dataset is small and gradual so a VTL stores one full compressed back + diffs for incremental/differential/full backups. Also, VTLs look for redundant data across servers; 10 similar linux servers will have the almost identical binaries.
I am currently looking at http://www.datadomain.com/ VTL to replace a 72 slot dual drive LTO 1st gen library.
A VTL costs a bit less than a regular tape library + all the tapes you need but the increased throughput and no more tape handling is what makes it worth it.
0.02cents.
Please note
940 is DDR Dual channel registered memory with ECC support
vs
939 is DDR Dual channel non-registered memory
Reference: http://www.hardocp.com/article.html?art=NjI2
One was designed for server/workstation requirements; registered and ECC support.
The other was designed for desktop requirements where price is and total system costs (ECC memory is significantly more expensive) are more important.
Please note I meant, Redhat patches don't break Redhat Applications vs Sun patches breaking SunOne Messaging server.
Thanks for pointing out sun's patchmanager.
I have yet to see benchmarks comparing Linux2.6, Solaris10 with BSD variants. Anyone have any links?
Most Oracle installations aren't run on Solaris..
a) Granted larger Oracle installations (8Processors or more) are on SPARC/Solaris.
b) Many small to medium sized installations are run on x86/Linux. Has been this way for a few years now, ever since Oracle started supporting Linux really aggressively.
Solaris's major advantage is standardized kernel, kernel APIs and system libraries.
It allows application developers to better target the platform they want to develop on and support and for how long. In the commercial space this is a big advantage for Solaris.
Where Solaris fails compared to say Redhat (note I am talking about commercial version) is how easy it is to manage the system.
Want to apply the latest patches that have been approved by Redhat?
up2date
For Solaris?
go to sun's site hunt for the right page that will list the latest patch cluster.
verify this patch cluster doesn't break any of your Sun applications (e.g. SunOne messaging 5.2 sp2 has problems with the latest patch cluster for solaris 9).
Why should a user have to hunt for this information?
Why should I have to phone support and have them hunt for it?
Why isn't this information on the patch cluster's download page?
Why doesn't Solaris have a patch management system that covers all Sun products installed on a server?
0.02c
US Gas consumption has steadly gone up for the past 10 yrs.
The reason prices are rising is increased demand from other nations, e.g. China (you know the country that makes all those cheap goods you can buy at walmart et al).
Never heard about this, do you have a source?
I wish I had mod points.
More I wish most americans understood and practiced this ideal.
If america treats others as equals then others in the world wouldn't resent them as much.
But I wonder if the american psyche can handle treating others as equals (the lack of superiority)?
Please excuse my ignorance but when did Iran publicly state it reserved the "First strike" right?
Everything I have read suggests they reserve the right to defend themselves and would attack another country only if they were attacked first.
Sounds pretty reasonable, especially compared to North Korea.
Please correct me if I am wrong.
For an Opteron board to go past 16GB, you need a quad opetron board + Opteron 800 series CPUs.
Opteron 800 series CPUs are much more expensive http://www.pricewatch.com/h/mn.aspx?i=3&f=1
Which brand on motherboard was this nforce2 board?
YMMV but cheap manufacturer's cut corners on board design, caps, etc using the same chipset, just sell a motherboard that is $50 cheaper than the rest.
Again YMMV.
Raja as a title translates to king.
In the indian subcontinent, it applied to lords of states/provinces, e.g. Raja of Hyderabad
Where Hyderabad is now a state in India.
I don't have stats for innocent deaths caused by either side. I am sure you could find some if one searched hard enough.
The link doesn't work for me.
.au only link?
Is it a
Your version of history isn't entirely correct/complete..
T ype=0& searchIn=0&query=kashmir&start=1&num=10&max=25&sor tBy=date
1. Yes the raja (roughly translates to king but is more like a local lord) of kashmir wanted to maintain his monarchy (didn't want either pakistan or india as both's political systems were shaping up to be purely parliamentary, i.e. no monarchy).
2. Under the terms of partition (agreed upon by muslim and hindu political parties..
a) in states where a popular majority (muslim or hindu) exists, that decides which country that state joins (geographical limitations allowing).
b) in a state where no clear majority exists, the ruler of the state will decide.
3. In the case of Kashmir..
a) there was and still is a muslim majority.
b) when the political reality of keeping kashmir as an independent monarchy wasn't possible, the raja choose india (against partition rules).
That is one of the major seeds of discontent that has lead to two wars between the two countries. All this is history.
For the last 20 years, in an attempt to stamp out terrorism as an excuse, the indian army has been targeting any muslims who raise a political voice; arrest, torture without plausible cause (sometimes leading to dead). It has only created a new generation of terrorists (freedom fighters) and only leads to lend weight to the calls for seperation from india.
I can only dig up one reference right now:
http://www.amnesty.org/results/is/eng?query
university scholarship:
You think your 1980's university fees covered the cost of the education you got?
Goverment transfer payments covered over 95% of the costs (of the university) for you.
Don't you think you have an obligation to pay for the next generation?
Just like the previous generation did?
You sir are trolling.
I know of two colleges in the lower mainland of BC, Canada that are using NetMail in production for the last 4years (provide email to all students). One of these colleges just bought a new portal system that comes bundled with SunOne messaging server (email integrated portal) and they still stuck with NetMail for email. Why? Out of the box it is designed with features that make an admin's life easier (think seamless email quotas, etc).
w.r.t to the rest of your comments about Sun..
They still stick to an attitude and culture that is elitiest and down right snotty.
Reminds me of IBM in the 70s.
I don't know if Sun will survive; they claim they want to be a services company but still want to sell hardware+software bundles while VARs provide the real services.
Specifically thinking about when the HW intermittently hard locks, have seen it happen on a server that I build from desktop class parts. Heck I have seen it happen on a G4 XServe (turned out to be bad ram; for 4wks server would cause the clients to crash intermittently).
Yes best to junk it but what if the cause is a defect in the drivers or worse in the hardware? Then any same model spare is useless too
At work we used to buy OEM SuperMicro servers from a local reseller for PC class servers.
Unix was mostly AIX with a bit of Sun Sparc with support provided by IBM or a local Sun reseller.
For PC class servers we initially used Mylex RAID controllers only. They were great boards and great RAID controllers. But a few problems..
a) supermicro server's have a very select list of compatible memory (vendor/model/batch). This added significantly to the cost.
b) Mylex RAID controllers can be temperamental to SCSI timing issues and their driver support does NOT expose RAID events (disk failure) to the host OS.
c) Mylex driver support for OSes other than Windows wasn't the greatest, driver's took many months to come out and usually required a firmware upgrade of the RAID array at the same time. This is fine on a new server that doesn't have any data but it is a big deal when you have a in production server with loads of data. If something screws up in the firmware update you can loose the RAID array, have to restore your data from backup which for large servers means at least a day of down time.
Also, at the time we were buying one class of server for all our server needs.
So we decided to switch to two class of PC servers..
1. Large: Dual/Quad Xeon SuperMicro Servers with Promise Ultratrak RM8000 external RAID arrays.
2. Small: Single P4 + Kingston ECC RAM + Promise IDE software RAID (RAID1) from a local whitebox reseller.
This worked well for a while except..
1. Large Servers:
a) Promise Ultratrak would hard lock when installing Redhat Enterprise 2.1. Something to do with Redhat tweaking the SCSI timings of the onboard SCSI controller (connects to external RAID) to suppport clustering. This was fixed by Promise with a firmware update about 14months after we encountered this problem.
b) WesternDigital's Accoustic Management would cause the Promise to timeout a drive causing the array to go critical, causing a rebuild to the hotspare, slowing the array down (50%) for 10+hrs (RAID 1+0 of eight 200GB disks). This was fixed about 8months later with a fix from western Digital.
c) The PRomise RAID array would not properly handle failed cached writes to disk, causing either corruption or drive to be mark bad (instead of marking the block bad and remapping it). Without disk caching enabled the performance of the Promise RAID array sucks.
2. Small:
a) Promise IDE RAID card (software RAID) was not supported well on most OSes except Windows. Their Linux driver left a lot to be desired.
b) One time we installed an Adaptec 2940UW scsi card (something we had lying around) to attach a local tape drive. Initially it worked great for a few months then the card started to die; the server would intermittenly reboot without any cause. We lived with this for a few months (all the time riding the local reseller's ass) until finally the card die gloriously, taking the motherboard and power supply with it. I have to give props to the reseller (Cybex Systems Inc. in Vancouver, BC, Canada), they were more than understanding and replaced everything in ASAP (sameday).
After all these problems it was decided we needed a top-tier name brand for our servers. After an RFP we decided on Dell; they were the cheapest by 20%, a local parts depot (this is important), their 4hr warranty was target for completed repair not just response (what HP/IBM have) and most importantly other companies provided a favourable recommendation.
Has it been great?
Well not exactly..
Two of the first few servers we purchased were for a two node Linux webserver setup where replication software would keep the two server's filesystems insync. On the Dell PowerEdge 2650, the embedded RAID controller is an Adaptec (some variation of single channel Ultra160 family). Unfortunately, adaptec controller's have many hardware defects that the driver has to compenstate for (most hard is like this, some more than others). With the replication software installed, within a