We are currently running UniData 8.2 on AIX 7.1 on soon to be Power10 IBM servers. We have roughly 2000 connections across "4" UniData instances linked by NFS and NFA data pointers sharing a pool of 22 Power server cpu running at smt8.
Our management wants to look into moving to a cheaper Linux environment and I am trying to gauge how good or bad of an idea that is and what that would look like vs our AIX setup we have been running for 25+ years.
We also have over 2000 AIX user accounts on each system, several hundred AIX printer queues, thousands of programming calls using and parsing output from various AIX commands.
If we drop this on Linux, what programming changes would need done to make the code run, outside of all the AIX vs Linux OS command vetting?
If you have a large UniData environment, I'd love to get on the phone and chat if anyone is up to it? Maybe even do a meet and great and see how your environment looks vs ours. Some knowledge sharing.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
Page 1 / 1
We are currently running UniData 8.2 on AIX 7.1 on soon to be Power10 IBM servers. We have roughly 2000 connections across "4" UniData instances linked by NFS and NFA data pointers sharing a pool of 22 Power server cpu running at smt8.
Our management wants to look into moving to a cheaper Linux environment and I am trying to gauge how good or bad of an idea that is and what that would look like vs our AIX setup we have been running for 25+ years.
We also have over 2000 AIX user accounts on each system, several hundred AIX printer queues, thousands of programming calls using and parsing output from various AIX commands.
If we drop this on Linux, what programming changes would need done to make the code run, outside of all the AIX vs Linux OS command vetting?
If you have a large UniData environment, I'd love to get on the phone and chat if anyone is up to it? Maybe even do a meet and great and see how your environment looks vs ours. Some knowledge sharing.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
Our management wants to look into moving to a cheaper Linux environment and I am trying to gauge how good or bad of an idea that is and what that would look like vs our AIX setup we have been running for 25+ years.
We also have over 2000 AIX user accounts on each system, several hundred AIX printer queues, thousands of programming calls using and parsing output from various AIX commands.
If we drop this on Linux, what programming changes would need done to make the code run, outside of all the AIX vs Linux OS command vetting?
If you have a large UniData environment, I'd love to get on the phone and chat if anyone is up to it? Maybe even do a meet and great and see how your environment looks vs ours. Some knowledge sharing.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
Hi Tom!
We're a Universe on AIX shop, and recently upgraded from P8s to P9s (project was started before P10s came out).
I would recommend contacting your IBM account team (if you haven't already). I'm sure you're not the first to come to them with the AIX vs. Linux question. Depending on who you work with, they may not have much info specific the the Multi-Value world, but should be able to provide you with the more generic info - comparative user counts, performance comparisons, maintenance and security comparisons, etc.
If you'd like to talk system and application configuration details, feel free to hit me up privately. Our InfoSec team frowns on us posting that level of information in public forums - and rightly so.
------------------------------
Brian Paige
------------------------------
We are currently running UniData 8.2 on AIX 7.1 on soon to be Power10 IBM servers. We have roughly 2000 connections across "4" UniData instances linked by NFS and NFA data pointers sharing a pool of 22 Power server cpu running at smt8.
Our management wants to look into moving to a cheaper Linux environment and I am trying to gauge how good or bad of an idea that is and what that would look like vs our AIX setup we have been running for 25+ years.
We also have over 2000 AIX user accounts on each system, several hundred AIX printer queues, thousands of programming calls using and parsing output from various AIX commands.
If we drop this on Linux, what programming changes would need done to make the code run, outside of all the AIX vs Linux OS command vetting?
If you have a large UniData environment, I'd love to get on the phone and chat if anyone is up to it? Maybe even do a meet and great and see how your environment looks vs ours. Some knowledge sharing.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
Our management wants to look into moving to a cheaper Linux environment and I am trying to gauge how good or bad of an idea that is and what that would look like vs our AIX setup we have been running for 25+ years.
We also have over 2000 AIX user accounts on each system, several hundred AIX printer queues, thousands of programming calls using and parsing output from various AIX commands.
If we drop this on Linux, what programming changes would need done to make the code run, outside of all the AIX vs Linux OS command vetting?
If you have a large UniData environment, I'd love to get on the phone and chat if anyone is up to it? Maybe even do a meet and great and see how your environment looks vs ours. Some knowledge sharing.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
Everything should drop on pretty seamlessly barring AIX / LINUX shell command variations as you have already commented, The areas where I would want to look closely at potential differences are:
- AIX/LINUX (CUPS?) print subsystem scalability
- LINUX kernel parameters - AIX is dynamic I seem to recall whereas LINUX is statically tuneable.
- File system performance and throughput. With AIX you are probably using JFS2 - ideally with j2_nPagesPerWriteBehindCluster set to zero - File system type and configuration on LINUX is quite different, in imposing queue limitations you may unconsciously be living with or have accommodated, These will not be relevant to LINUX
- JFS2 and LVM on AIX have some quirks
- NFS - I have encountered difficulties with NFS4 ACL compatibility I could not resolve and reverted to NFS3 in some non-UniData situations. If planning on using NFS with ACLs this is worth a check at least.
- User authentication throughput under peak busy load. I am not sure whether you will be continuing to use LDAP or moving to some other domain-based authentication, but the speed of the authentication server(s) and DNS lookup can be critical to performance.
- SCSI? If so then the usual precautions regarding which devices go on which SCSI bus. With AIX the IBM engineers building a system usually take care of this for you.
- ZFS and ZPOOLs? - Definitely worth looking at
- SDLC SSD drives recommended.
A full benchmark under peak-busy load for batch and interactive workloads is strongly recommended. I would set an expectation of a number of iterations with changes to the environment so that an optimal configuration can be determined. If a large hardware sale sits behind this then the hardware vendoe may be able to offer assistance from their Lab. in benchmarking (e.g. IBM or HP Labs,) with SE assistance.)
I'll leave it up to others for direct field experience beyond benchmarking.
Regards
JJ
------------------------------
John Jenkins
Thame, Oxfordshire
------------------------------
Hi Tom!
We're a Universe on AIX shop, and recently upgraded from P8s to P9s (project was started before P10s came out).
I would recommend contacting your IBM account team (if you haven't already). I'm sure you're not the first to come to them with the AIX vs. Linux question. Depending on who you work with, they may not have much info specific the the Multi-Value world, but should be able to provide you with the more generic info - comparative user counts, performance comparisons, maintenance and security comparisons, etc.
If you'd like to talk system and application configuration details, feel free to hit me up privately. Our InfoSec team frowns on us posting that level of information in public forums - and rightly so.
------------------------------
Brian Paige
------------------------------
The usual issue that comes up inn this area is that the disk subsystem organisation and file system type should be tuned for a large volume of relatively small random access reads and writes. Often the 'default' is for a large SQL-type database with very large streamed I/O reads and writes ad it gets missed by vendors.
That the AIX setting for j2_nPagesPerWriteBehindCluster should be changed from the default 32 to zero is just one very specific example..
Regards
JJ
------------------------------
John Jenkins
Thame, Oxfordshire
------------------------------
Brian,
The usual issue that comes up inn this area is that the disk subsystem organisation and file system type should be tuned for a large volume of relatively small random access reads and writes. Often the 'default' is for a large SQL-type database with very large streamed I/O reads and writes ad it gets missed by vendors.
That the AIX setting for j2_nPagesPerWriteBehindCluster should be changed from the default 32 to zero is just one very specific example..
Regards
JJ
------------------------------
John Jenkins
Thame, Oxfordshire
------------------------------
The usual issue that comes up inn this area is that the disk subsystem organisation and file system type should be tuned for a large volume of relatively small random access reads and writes. Often the 'default' is for a large SQL-type database with very large streamed I/O reads and writes ad it gets missed by vendors.
That the AIX setting for j2_nPagesPerWriteBehindCluster should be changed from the default 32 to zero is just one very specific example..
Regards
JJ
------------------------------
John Jenkins
Thame, Oxfordshire
------------------------------
AND, I've been kind of spoiled as far as SAN disk goes. Our SAN team explained to me that once the AIX settings are in place, there were very few settings on the SAN side since it auto-optimizes for the type of access that the database and applications perform. Having to think about that has left my brain for the most part.
My recommendation for talking to IBM had to do with just getting the base hardware equivalencies set up - how a Linux box (or multiple Linux boxes) would have to be configured in order to provide comparable performance to the IBM systems already in use.
Brian
Tom,
Everything should drop on pretty seamlessly barring AIX / LINUX shell command variations as you have already commented, The areas where I would want to look closely at potential differences are:
A full benchmark under peak-busy load for batch and interactive workloads is strongly recommended. I would set an expectation of a number of iterations with changes to the environment so that an optimal configuration can be determined. If a large hardware sale sits behind this then the hardware vendoe may be able to offer assistance from their Lab. in benchmarking (e.g. IBM or HP Labs,) with SE assistance.)
I'll leave it up to others for direct field experience beyond benchmarking.
Regards
JJ
------------------------------
John Jenkins
Thame, Oxfordshire
------------------------------
Everything should drop on pretty seamlessly barring AIX / LINUX shell command variations as you have already commented, The areas where I would want to look closely at potential differences are:
- AIX/LINUX (CUPS?) print subsystem scalability
- LINUX kernel parameters - AIX is dynamic I seem to recall whereas LINUX is statically tuneable.
- File system performance and throughput. With AIX you are probably using JFS2 - ideally with j2_nPagesPerWriteBehindCluster set to zero - File system type and configuration on LINUX is quite different, in imposing queue limitations you may unconsciously be living with or have accommodated, These will not be relevant to LINUX
- JFS2 and LVM on AIX have some quirks
- NFS - I have encountered difficulties with NFS4 ACL compatibility I could not resolve and reverted to NFS3 in some non-UniData situations. If planning on using NFS with ACLs this is worth a check at least.
- User authentication throughput under peak busy load. I am not sure whether you will be continuing to use LDAP or moving to some other domain-based authentication, but the speed of the authentication server(s) and DNS lookup can be critical to performance.
- SCSI? If so then the usual precautions regarding which devices go on which SCSI bus. With AIX the IBM engineers building a system usually take care of this for you.
- ZFS and ZPOOLs? - Definitely worth looking at
- SDLC SSD drives recommended.
A full benchmark under peak-busy load for batch and interactive workloads is strongly recommended. I would set an expectation of a number of iterations with changes to the environment so that an optimal configuration can be determined. If a large hardware sale sits behind this then the hardware vendoe may be able to offer assistance from their Lab. in benchmarking (e.g. IBM or HP Labs,) with SE assistance.)
I'll leave it up to others for direct field experience beyond benchmarking.
Regards
JJ
------------------------------
John Jenkins
Thame, Oxfordshire
------------------------------
* All files will have to be converted with fnuxi
* Some/All programs may need to be recompiled depending on the options under which they were originally compiled
* If the software is vendor-provided, the vendor may need to provide a different version of the object code, assuming they support the Linux environment
* I don't remember if fnuxi works on object code
* Review the UVCONFIG file, as some of the settings are different between the types of systems, and some only apply to specific environments
* Check bug/hotfix readme files relevant to your release level - some fixes only apply to specific environments
* For that matter, check all release documentation to ensure you're not using something that's currently available in AIX, but not yet available in Linux (or to see if there's something available in Linux that you would now be able to use)
* When someone throws out that 'cheaper' phrase, sometimes there are costs that are not considered, above and beyond the initial conversion - retraining of staff, system stability, system security, etc. Doing a cost-risk analysis and attempting to identity the costs associated with the risks of each environment is something that tends to be overlooked - or given only minimum effort.
Brian
Tom,
Everything should drop on pretty seamlessly barring AIX / LINUX shell command variations as you have already commented, The areas where I would want to look closely at potential differences are:
A full benchmark under peak-busy load for batch and interactive workloads is strongly recommended. I would set an expectation of a number of iterations with changes to the environment so that an optimal configuration can be determined. If a large hardware sale sits behind this then the hardware vendoe may be able to offer assistance from their Lab. in benchmarking (e.g. IBM or HP Labs,) with SE assistance.)
I'll leave it up to others for direct field experience beyond benchmarking.
Regards
JJ
------------------------------
John Jenkins
Thame, Oxfordshire
------------------------------
Everything should drop on pretty seamlessly barring AIX / LINUX shell command variations as you have already commented, The areas where I would want to look closely at potential differences are:
- AIX/LINUX (CUPS?) print subsystem scalability
- LINUX kernel parameters - AIX is dynamic I seem to recall whereas LINUX is statically tuneable.
- File system performance and throughput. With AIX you are probably using JFS2 - ideally with j2_nPagesPerWriteBehindCluster set to zero - File system type and configuration on LINUX is quite different, in imposing queue limitations you may unconsciously be living with or have accommodated, These will not be relevant to LINUX
- JFS2 and LVM on AIX have some quirks
- NFS - I have encountered difficulties with NFS4 ACL compatibility I could not resolve and reverted to NFS3 in some non-UniData situations. If planning on using NFS with ACLs this is worth a check at least.
- User authentication throughput under peak busy load. I am not sure whether you will be continuing to use LDAP or moving to some other domain-based authentication, but the speed of the authentication server(s) and DNS lookup can be critical to performance.
- SCSI? If so then the usual precautions regarding which devices go on which SCSI bus. With AIX the IBM engineers building a system usually take care of this for you.
- ZFS and ZPOOLs? - Definitely worth looking at
- SDLC SSD drives recommended.
A full benchmark under peak-busy load for batch and interactive workloads is strongly recommended. I would set an expectation of a number of iterations with changes to the environment so that an optimal configuration can be determined. If a large hardware sale sits behind this then the hardware vendoe may be able to offer assistance from their Lab. in benchmarking (e.g. IBM or HP Labs,) with SE assistance.)
I'll leave it up to others for direct field experience beyond benchmarking.
Regards
JJ
------------------------------
John Jenkins
Thame, Oxfordshire
------------------------------
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
Other things to consider:
* All files will have to be converted with fnuxi
* Some/All programs may need to be recompiled depending on the options under which they were originally compiled
* If the software is vendor-provided, the vendor may need to provide a different version of the object code, assuming they support the Linux environment
* I don't remember if fnuxi works on object code
* Review the UVCONFIG file, as some of the settings are different between the types of systems, and some only apply to specific environments
* Check bug/hotfix readme files relevant to your release level - some fixes only apply to specific environments
* For that matter, check all release documentation to ensure you're not using something that's currently available in AIX, but not yet available in Linux (or to see if there's something available in Linux that you would now be able to use)
* When someone throws out that 'cheaper' phrase, sometimes there are costs that are not considered, above and beyond the initial conversion - retraining of staff, system stability, system security, etc. Doing a cost-risk analysis and attempting to identity the costs associated with the risks of each environment is something that tends to be overlooked - or given only minimum effort.
Brian
* All files will have to be converted with fnuxi
* Some/All programs may need to be recompiled depending on the options under which they were originally compiled
* If the software is vendor-provided, the vendor may need to provide a different version of the object code, assuming they support the Linux environment
* I don't remember if fnuxi works on object code
* Review the UVCONFIG file, as some of the settings are different between the types of systems, and some only apply to specific environments
* Check bug/hotfix readme files relevant to your release level - some fixes only apply to specific environments
* For that matter, check all release documentation to ensure you're not using something that's currently available in AIX, but not yet available in Linux (or to see if there's something available in Linux that you would now be able to use)
* When someone throws out that 'cheaper' phrase, sometimes there are costs that are not considered, above and beyond the initial conversion - retraining of staff, system stability, system security, etc. Doing a cost-risk analysis and attempting to identity the costs associated with the risks of each environment is something that tends to be overlooked - or given only minimum effort.
Brian
I am basically 99% sure we are running the best of breed for UniData in our environment, which is quite complex vs other environments, but I am looking for data to change my mind. I do not work for IBM, I work for my company so that's where my mind set is. I am going to compile all the pros and cons, pitfalls, knowns and unknowns that I can and if management still wants to trudge on we will attempt to benchmark. I just do not see how AIX on IBM Power10 with flash drives can be out performed or out scaled by Linux running Intel based hardware at a true cost savings, definitely not on the same footprint. Change my mind, lol.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
We are currently running UniData 8.2 on AIX 7.1 on soon to be Power10 IBM servers. We have roughly 2000 connections across "4" UniData instances linked by NFS and NFA data pointers sharing a pool of 22 Power server cpu running at smt8.
Our management wants to look into moving to a cheaper Linux environment and I am trying to gauge how good or bad of an idea that is and what that would look like vs our AIX setup we have been running for 25+ years.
We also have over 2000 AIX user accounts on each system, several hundred AIX printer queues, thousands of programming calls using and parsing output from various AIX commands.
If we drop this on Linux, what programming changes would need done to make the code run, outside of all the AIX vs Linux OS command vetting?
If you have a large UniData environment, I'd love to get on the phone and chat if anyone is up to it? Maybe even do a meet and great and see how your environment looks vs ours. Some knowledge sharing.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
Our management wants to look into moving to a cheaper Linux environment and I am trying to gauge how good or bad of an idea that is and what that would look like vs our AIX setup we have been running for 25+ years.
We also have over 2000 AIX user accounts on each system, several hundred AIX printer queues, thousands of programming calls using and parsing output from various AIX commands.
If we drop this on Linux, what programming changes would need done to make the code run, outside of all the AIX vs Linux OS command vetting?
If you have a large UniData environment, I'd love to get on the phone and chat if anyone is up to it? Maybe even do a meet and great and see how your environment looks vs ours. Some knowledge sharing.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
Using SSL?
- Any and all server certificates you use will need to be recreated,
- Changing domain? Any cross-domain trust relationships will need to be amended,
Check both user and global environment variable settings on the AIX server as these can sometimes be missed in a migration,
Any web servers or SB/XA servers involved? These will need changes as well.
Using cloud or iron? Each has its own considerations, and remember that 'cloud' is just another term for 'Someone else's computer'.
External software involved? If so, check availability and potential costs on LINUX (ZIp/post code name and address lookup software is the most common - linked via CALLC).
Implementation:
* Parallel run.... please.... this can save so much grief.
Regards & Blwyddyn Newydd Dda
JJ
------------------------------
John Jenkins
Thame, Oxfordshire
------------------------------
We are currently running UniData 8.2 on AIX 7.1 on soon to be Power10 IBM servers. We have roughly 2000 connections across "4" UniData instances linked by NFS and NFA data pointers sharing a pool of 22 Power server cpu running at smt8.
Our management wants to look into moving to a cheaper Linux environment and I am trying to gauge how good or bad of an idea that is and what that would look like vs our AIX setup we have been running for 25+ years.
We also have over 2000 AIX user accounts on each system, several hundred AIX printer queues, thousands of programming calls using and parsing output from various AIX commands.
If we drop this on Linux, what programming changes would need done to make the code run, outside of all the AIX vs Linux OS command vetting?
If you have a large UniData environment, I'd love to get on the phone and chat if anyone is up to it? Maybe even do a meet and great and see how your environment looks vs ours. Some knowledge sharing.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
Our management wants to look into moving to a cheaper Linux environment and I am trying to gauge how good or bad of an idea that is and what that would look like vs our AIX setup we have been running for 25+ years.
We also have over 2000 AIX user accounts on each system, several hundred AIX printer queues, thousands of programming calls using and parsing output from various AIX commands.
If we drop this on Linux, what programming changes would need done to make the code run, outside of all the AIX vs Linux OS command vetting?
If you have a large UniData environment, I'd love to get on the phone and chat if anyone is up to it? Maybe even do a meet and great and see how your environment looks vs ours. Some knowledge sharing.
------------------------------
Tom Vankirk
Unix admin
Cabinetworks Group Michigan, LLC
Garrettsville OH US
------------------------------
I suggest you try a dry run, moving some accounts and data to a Linux server.
For the most part, the advice provided for a UniVerse migration does apply to UniData, with the obvious differences in commands and configuration parameters/files.
Note that I believe you identified the main issues in moving between OS and Hardware. The OS commands you execute and parse will need to be identified and modified for the new OS. In your planning I suggest you include ample time to test the solution before going into production.
------------------------------
Mike Rajkowski
MultiValue Product Evangelist
Rocket Internal - All Brands
DENVER CO US
------------------------------
Tom,
I suggest you try a dry run, moving some accounts and data to a Linux server.
For the most part, the advice provided for a UniVerse migration does apply to UniData, with the obvious differences in commands and configuration parameters/files.
Note that I believe you identified the main issues in moving between OS and Hardware. The OS commands you execute and parse will need to be identified and modified for the new OS. In your planning I suggest you include ample time to test the solution before going into production.
------------------------------
Mike Rajkowski
MultiValue Product Evangelist
Rocket Internal - All Brands
DENVER CO US
------------------------------
I suggest you try a dry run, moving some accounts and data to a Linux server.
For the most part, the advice provided for a UniVerse migration does apply to UniData, with the obvious differences in commands and configuration parameters/files.
Note that I believe you identified the main issues in moving between OS and Hardware. The OS commands you execute and parse will need to be identified and modified for the new OS. In your planning I suggest you include ample time to test the solution before going into production.
------------------------------
Mike Rajkowski
MultiValue Product Evangelist
Rocket Internal - All Brands
DENVER CO US
------------------------------
We have just successfully migrated a 25+ year, highly customized, legacy system from SunSolaris unix to RedHat 7, Universe 11.3.4. We hired Rocket Professional services, Alan Cornell did a fantastic job of configuring and navigating the Universe upgrade, our system admin, myself (a 30+ pick vetran) and a 25+ business knowledge expert from our staff worked only on the migration for 6+ months. Our business analyst tested constantly. We held UAT for 3 weeks, having users test their daily job functions. We had to "re-write" the entire login routine, as the company had old "C" code programs that did most of the user validation and launching of UV. Just to get past the old 'C' code took us 2 months, but then we seemed to move along quickly in getting data over, fnuxied, and dictionaries compiled, and programs which did OS commands modified for proper syntax. In August, we started weekly "dry-runs" of our migration tasks, keeping a log of what needed to be done, and worked on writing scripts and programs, to speed the process, from backing up the old system, restoring data , etc... Each person had dedicated tasks to do on go live and we strictly followed our schedule. Our company actually customized Universe code and commands over the years,which proved to be quite a challenge making sure we replicated the customizations. There are 22K accounts and 2.5M data files. Rocket was instrumental and we could not have done it without their support. I would highly encourage you to look into using their services . We came in under budget and on time.
Good luck to you, and feel free to reach out if you would like any details of our efforts
------------------------------
Angela Bolton
Product Owner
S & B Administrative Services, Ltd.
Houston TX US
acbolton@sbec.com
------------------------------
Hello,
We have just successfully migrated a 25+ year, highly customized, legacy system from SunSolaris unix to RedHat 7, Universe 11.3.4. We hired Rocket Professional services, Alan Cornell did a fantastic job of configuring and navigating the Universe upgrade, our system admin, myself (a 30+ pick vetran) and a 25+ business knowledge expert from our staff worked only on the migration for 6+ months. Our business analyst tested constantly. We held UAT for 3 weeks, having users test their daily job functions. We had to "re-write" the entire login routine, as the company had old "C" code programs that did most of the user validation and launching of UV. Just to get past the old 'C' code took us 2 months, but then we seemed to move along quickly in getting data over, fnuxied, and dictionaries compiled, and programs which did OS commands modified for proper syntax. In August, we started weekly "dry-runs" of our migration tasks, keeping a log of what needed to be done, and worked on writing scripts and programs, to speed the process, from backing up the old system, restoring data , etc... Each person had dedicated tasks to do on go live and we strictly followed our schedule. Our company actually customized Universe code and commands over the years,which proved to be quite a challenge making sure we replicated the customizations. There are 22K accounts and 2.5M data files. Rocket was instrumental and we could not have done it without their support. I would highly encourage you to look into using their services . We came in under budget and on time.
Good luck to you, and feel free to reach out if you would like any details of our efforts
------------------------------
Angela Bolton
Product Owner
S & B Administrative Services, Ltd.
Houston TX US
acbolton@sbec.com
------------------------------
We have just successfully migrated a 25+ year, highly customized, legacy system from SunSolaris unix to RedHat 7, Universe 11.3.4. We hired Rocket Professional services, Alan Cornell did a fantastic job of configuring and navigating the Universe upgrade, our system admin, myself (a 30+ pick vetran) and a 25+ business knowledge expert from our staff worked only on the migration for 6+ months. Our business analyst tested constantly. We held UAT for 3 weeks, having users test their daily job functions. We had to "re-write" the entire login routine, as the company had old "C" code programs that did most of the user validation and launching of UV. Just to get past the old 'C' code took us 2 months, but then we seemed to move along quickly in getting data over, fnuxied, and dictionaries compiled, and programs which did OS commands modified for proper syntax. In August, we started weekly "dry-runs" of our migration tasks, keeping a log of what needed to be done, and worked on writing scripts and programs, to speed the process, from backing up the old system, restoring data , etc... Each person had dedicated tasks to do on go live and we strictly followed our schedule. Our company actually customized Universe code and commands over the years,which proved to be quite a challenge making sure we replicated the customizations. There are 22K accounts and 2.5M data files. Rocket was instrumental and we could not have done it without their support. I would highly encourage you to look into using their services . We came in under budget and on time.
Good luck to you, and feel free to reach out if you would like any details of our efforts
------------------------------
Angela Bolton
Product Owner
S & B Administrative Services, Ltd.
Houston TX US
acbolton@sbec.com
------------------------------
22K accounts! Are those user logins?
------------------------------
Peter Gonzalez
Programmer/Systems Analyst/EDI Developer
Self Registered
St. Louis MO US
------------------------------
Hello Angela.
22K accounts! Are those user logins?
------------------------------
Peter Gonzalez
Programmer/Systems Analyst/EDI Developer
Self Registered
St. Louis MO US
------------------------------
22K accounts! Are those user logins?
------------------------------
Peter Gonzalez
Programmer/Systems Analyst/EDI Developer
Self Registered
St. Louis MO US
------------------------------
------------------------------
Angela Bolton
Product Owner
S & B Administrative Services, Ltd.
Houston TX US
acbolton@sbec.com
------------------------------
Sign up
Already have an account? Login
Welcome to the Rocket Forum!
Please log in or register:
Employee Login | Registration Member Login | RegistrationEnter your E-mail address. We'll send you an e-mail with instructions to reset your password.
77 4th Avenue
Waltham, MA 02451 USA
