HDR for Enterprise DR solution?

Hi, We are exploring using HDR / RSS for an enterprise DR (Disaster Recovery) solution. We are currently on 11.50.FC6 (eventually will upgrade to 11.70). We have a heavily used OLTP system with 2,000 + concurrent sessions. We are looking for a rock solid DR solution. Currently we use True-Copy disk replication from Site A to Site B and are looking to replace that with an Informix based DB solution. We will have dual 10 Gb lines between the sites with a distance of 110 km (69 Miles). I am told the latency will be very low - as if the 2 servers were in the same data center. We are using top of the line, fast Unix Solaris servers. Does anyone have any input? Is anyone running a similar production setup with fast servers, many concurrent sessions and using HDR ? Any real production based feedback would be greatly appreciated. Also feel free to email me offline. Thank You, David in4mixdba@gmail.com --20cf3066791994ed9304c3ed8765

HDR just works. As to whether the lines between the data centers will be sufficient to keep up, look at your peak logical log consumption rates and extrapolate from there. If that is less than the bandwidth of the lines then you are good to go. That's really the only bottleneck there can be. HDR will keep up on the secondary as long as the hardware on the secondary is at least as good as the primary's iron. I would run the secondary in asynchronous mode with zero apply delay and with index replication enabled. Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Tue, Jul 3, 2012 at 10:14 AM, Informix DBA <in4mixdba@gmail.com> wrote: > Hi, > > We are exploring using HDR / RSS for an enterprise DR (Disaster Recovery) > solution. We are currently on 11.50.FC6 (eventually will upgrade to > 11.70). We have a heavily used OLTP system with 2,000 + concurrent > sessions. We are looking for a rock solid DR solution. Currently we use > True-Copy disk replication from Site A to Site B and are looking to replace > that with an Informix based DB solution. We will have dual 10 Gb lines > between the sites with a distance of 110 km (69 Miles). I am told the > latency will be very low - as if the 2 servers were in the same data > center. We are using top of the line, fast Unix Solaris servers. > > Does anyone have any input? Is anyone running a similar production setup > with fast servers, many concurrent sessions and using HDR ? Any real > production based feedback would be greatly appreciated. Also feel free to > email me offline. > > Thank You, > > David > in4mixdba@gmail.com > > --20cf3066791994ed9304c3ed8765 > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --14dae9340e1b6fe24104c3edadc7

Not sure if I'll answer your exact question, because I don't know if you're concerned with the distance or the load. I work frequently on a customer site using HDR/RSS on the same datacenter. They reach 3000 sessions frequently. They have 1 HDR and 1 RSS. The machines are HP-UX with around 3/4 years. No problems. I used to go to another customer with an HDR around 300KM apart from the primary. Usually no problems but less load than the other customer. They had some issues if the connection went down, but this was V10... With 11.10+ you have the option to use RSS which should not have any impact on the primary (aprt the fact that the primary needs to send the logs) With HDR you may have some issues with latency during checkpoints. Using Informix DR can be cheaper than other disk based solutions. Nevertheless I also have customers using disk based because they use it for other systems and the size of the database is "small" (I believe the disk replication feature cost is based on size...?) Of course, you should test it... For HDR, the parameters can make a difference. And you have to evaluate if you can use it.... Are your databases logged? Do you usually use raw tables?... Regards. On Tue, Jul 3, 2012 at 3:14 PM, Informix DBA <in4mixdba@gmail.com> wrote: > Hi, > > We are exploring using HDR / RSS for an enterprise DR (Disaster Recovery) > solution. We are currently on 11.50.FC6 (eventually will upgrade to > 11.70). We have a heavily used OLTP system with 2,000 + concurrent > sessions. We are looking for a rock solid DR solution. Currently we use > True-Copy disk replication from Site A to Site B and are looking to replace > that with an Informix based DB solution. We will have dual 10 Gb lines > between the sites with a distance of 110 km (69 Miles). I am told the > latency will be very low - as if the 2 servers were in the same data > center. We are using top of the line, fast Unix Solaris servers. > > Does anyone have any input? Is anyone running a similar production setup > with fast servers, many concurrent sessions and using HDR ? Any real > production based feedback would be greatly appreciated. Also feel free to > email me offline. > > Thank You, > > David > in4mixdba@gmail.com > > --20cf3066791994ed9304c3ed8765 > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Fernando Nunes Portugal http://informix-technology.blogspot.com My email works... but I don't check it frequently... --20cf303b40ef03b0d004c3ee12da

Art, Thank you. The secondary hardware is identical to the primary. I recall reading that in the past HDR had issues both with the secondary crashing and with causing the primary to slow when the secondary gets out of sync. I believe there was thought to the effect that, "Processors have gotten much faster and that exposes certain issues in HDR"? I know we once tried working with HDR in async and after a planned failover to the secondary, the old primary - new secondary ended up crashing after a few hours. There was no resolution with that PMR. Any known issues with HDR in 11.50 or 11.70 ? Thank you, --David On Tue, Jul 3, 2012 at 10:25 AM, Art Kagel <art.kagel@gmail.com> wrote: > HDR just works. As to whether the lines between the data centers will be > sufficient to keep up, look at your peak logical log consumption rates and > extrapolate from there. If that is less than the bandwidth of the lines > then you are good to go. That's really the only bottleneck there can be. > HDR will keep up on the secondary as long as the hardware on the secondary > is at least as good as the primary's iron. I would run the secondary in > asynchronous mode with zero apply delay and with index replication enabled. > > Art > > Art S. Kagel > Advanced DataTools (www.advancedatatools.com) > Blog: http://informix-myview.blogspot.com/ > > Disclaimer: Please keep in mind that my own opinions are my own opinions > and do not reflect on my employer, Advanced DataTools, the IIUG, nor any > other organization with which I am associated either explicitly, > implicitly, or by inference. Neither do those opinions reflect those of > other individuals affiliated with any entity with which I am affiliated nor > those of the entities themselves. > > On Tue, Jul 3, 2012 at 10:14 AM, Informix DBA <in4mixdba@gmail.com> wrote: > > > Hi, > > > > We are exploring using HDR / RSS for an enterprise DR (Disaster Recovery) > > solution. We are currently on 11.50.FC6 (eventually will upgrade to > > 11.70). We have a heavily used OLTP system with 2,000 + concurrent > > sessions. We are looking for a rock solid DR solution. Currently we use > > True-Copy disk replication from Site A to Site B and are looking to > replace > > that with an Informix based DB solution. We will have dual 10 Gb lines > > between the sites with a distance of 110 km (69 Miles). I am told the > > latency will be very low - as if the 2 servers were in the same data > > center. We are using top of the line, fast Unix Solaris servers. > > > > Does anyone have any input? Is anyone running a similar production setup > > with fast servers, many concurrent sessions and using HDR ? Any real > > production based feedback would be greatly appreciated. Also feel free to > > email me offline. > > > > Thank You, > > > > David > > in4mixdba@gmail.com > > > > --20cf3066791994ed9304c3ed8765 > > > > > > > > > > ******************************************************************************* > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > > > --14dae9340e1b6fe24104c3edadc7 > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --20cf306679195a3f7404c3eed8ec

None that I am aware of. We have HDR working without problems at several clients in 11.50, 11.70, 10.00, and 7.31 Art Art S. Kagel Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Tue, Jul 3, 2012 at 11:48 AM, Informix DBA <in4mixdba@gmail.com> wrote: > Art, > > Thank you. The secondary hardware is identical to the primary. I recall > reading that in the past HDR had issues both with the secondary crashing > and with causing the primary to slow when the secondary gets out of sync. > I believe there was thought to the effect that, "Processors have gotten > much faster and that exposes certain issues in HDR"? I know we once tried > working with HDR in async and after a planned failover to the secondary, > the old primary - new secondary ended up crashing after a few hours. There > was no resolution with that PMR. > > Any known issues with HDR in 11.50 or 11.70 ? > > Thank you, > > --David > > On Tue, Jul 3, 2012 at 10:25 AM, Art Kagel <art.kagel@gmail.com> wrote: > > > HDR just works. As to whether the lines between the data centers will be > > sufficient to keep up, look at your peak logical log consumption rates > and > > extrapolate from there. If that is less than the bandwidth of the lines > > then you are good to go. That's really the only bottleneck there can be. > > HDR will keep up on the secondary as long as the hardware on the > secondary > > is at least as good as the primary's iron. I would run the secondary in > > asynchronous mode with zero apply delay and with index replication > enabled. > > > > Art > > > > Art S. Kagel > > Advanced DataTools (www.advancedatatools.com) > > Blog: http://informix-myview.blogspot.com/ > > > > Disclaimer: Please keep in mind that my own opinions are my own opinions > > and do not reflect on my employer, Advanced DataTools, the IIUG, nor any > > other organization with which I am associated either explicitly, > > implicitly, or by inference. Neither do those opinions reflect those of > > other individuals affiliated with any entity with which I am affiliated > nor > > those of the entities themselves. > > > > On Tue, Jul 3, 2012 at 10:14 AM, Informix DBA <in4mixdba@gmail.com> > wrote: > > > > > Hi, > > > > > > We are exploring using HDR / RSS for an enterprise DR (Disaster > Recovery) > > > solution. We are currently on 11.50.FC6 (eventually will upgrade to > > > 11.70). We have a heavily used OLTP system with 2,000 + concurrent > > > sessions. We are looking for a rock solid DR solution. Currently we use > > > True-Copy disk replication from Site A to Site B and are looking to > > replace > > > that with an Informix based DB solution. We will have dual 10 Gb lines > > > between the sites with a distance of 110 km (69 Miles). I am told the > > > latency will be very low - as if the 2 servers were in the same data > > > center. We are using top of the line, fast Unix Solaris servers. > > > > > > Does anyone have any input? Is anyone running a similar production > setup > > > with fast servers, many concurrent sessions and using HDR ? Any real > > > production based feedback would be greatly appreciated. Also feel free > to > > > email me offline. > > > > > > Thank You, > > > > > > David > > > in4mixdba@gmail.com > > > > > > --20cf3066791994ed9304c3ed8765 > > > > > > > > > > > > > > > > > > ******************************************************************************* > > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > > > > > > > --14dae9340e1b6fe24104c3edadc7 > > > > > > > > > > ******************************************************************************* > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > > > --20cf306679195a3f7404c3eed8ec > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --14dae9340e1b23213e04c3eee594

We have been using HDR since version 7, and replication has come a long way with version 11. We are currently on 11.50 and it works wonderfully. I hear that 11.70 is even better. > To: ids@iiug.org > From: domusonline@gmail.com > Subject: Re: HDR for Enterprise DR solution? [27525] > Date: Tue, 3 Jul 2012 10:53:03 -0400 > > Not sure if I'll answer your exact question, because I don't know if you're > concerned with the distance or the load. > I work frequently on a customer site using HDR/RSS on the same datacenter. > They reach 3000 sessions frequently. > They have 1 HDR and 1 RSS. The machines are HP-UX with around 3/4 years. No > problems. > > I used to go to another customer with an HDR around 300KM apart from the > primary. Usually no problems but less load than the other customer. > They had some issues if the connection went down, but this was V10... > > With 11.10+ you have the option to use RSS which should not have any impact > on the primary (aprt the fact that the primary needs to send the logs) > With HDR you may have some issues with latency during checkpoints. > > Using Informix DR can be cheaper than other disk based solutions. > Nevertheless I also have customers using disk based because they use it for > other systems and the size of the database is "small" (I believe the disk > replication feature cost is based on size...?) > > Of course, you should test it... For HDR, the parameters can make a > difference. > And you have to evaluate if you can use it.... Are your databases logged? > Do you usually use raw tables?... > > Regards. > > On Tue, Jul 3, 2012 at 3:14 PM, Informix DBA <in4mixdba@gmail.com> wrote: > > > Hi, > > > > We are exploring using HDR / RSS for an enterprise DR (Disaster Recovery) > > solution. We are currently on 11.50.FC6 (eventually will upgrade to > > 11.70). We have a heavily used OLTP system with 2,000 + concurrent > > sessions. We are looking for a rock solid DR solution. Currently we use > > True-Copy disk replication from Site A to Site B and are looking to replace > > that with an Informix based DB solution. We will have dual 10 Gb lines > > between the sites with a distance of 110 km (69 Miles). I am told the > > latency will be very low - as if the 2 servers were in the same data > > center. We are using top of the line, fast Unix Solaris servers. > > > > Does anyone have any input? Is anyone running a similar production setup > > with fast servers, many concurrent sessions and using HDR ? Any real > > production based feedback would be greatly appreciated. Also feel free to > > email me offline. > > > > Thank You, > > > > David > > in4mixdba@gmail.com > > > > --20cf3066791994ed9304c3ed8765 > > > > > > > > > ******************************************************************************* > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > > > -- > Fernando Nunes > Portugal > > http://informix-technology.blogspot.com > My email works... but I don't check it frequently... > > --20cf303b40ef03b0d004c3ee12da > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. >

Fernando, Thank you for your feedback. We are most concerned with the stability of the servers and replication i.e. not having servers crash. Our DB's are logged and have avoided using raw tables (IIRC). We are most interested in discovering if other customers are using the HDR / RSS as an enterprise solution with high-end hardware i.e. fast processors and a heavily used OLTP DB with 2,000 + concurrent sessions. Thank you, David On Tue, Jul 3, 2012 at 10:53 AM, Fernando Nunes <domusonline@gmail.com>wrote: > Not sure if I'll answer your exact question, because I don't know if you're > concerned with the distance or the load. > I work frequently on a customer site using HDR/RSS on the same datacenter. > They reach 3000 sessions frequently. > They have 1 HDR and 1 RSS. The machines are HP-UX with around 3/4 years. No > problems. > > I used to go to another customer with an HDR around 300KM apart from the > primary. Usually no problems but less load than the other customer. > They had some issues if the connection went down, but this was V10... > > With 11.10+ you have the option to use RSS which should not have any impact > on the primary (aprt the fact that the primary needs to send the logs) > With HDR you may have some issues with latency during checkpoints. > > Using Informix DR can be cheaper than other disk based solutions. > Nevertheless I also have customers using disk based because they use it for > other systems and the size of the database is "small" (I believe the disk > replication feature cost is based on size...?) > > Of course, you should test it... For HDR, the parameters can make a > difference. > And you have to evaluate if you can use it.... Are your databases logged? > Do you usually use raw tables?... > > Regards. > > On Tue, Jul 3, 2012 at 3:14 PM, Informix DBA <in4mixdba@gmail.com> wrote: > > > Hi, > > > > We are exploring using HDR / RSS for an enterprise DR (Disaster Recovery) > > solution. We are currently on 11.50.FC6 (eventually will upgrade to > > 11.70). We have a heavily used OLTP system with 2,000 + concurrent > > sessions. We are looking for a rock solid DR solution. Currently we use > > True-Copy disk replication from Site A to Site B and are looking to > replace > > that with an Informix based DB solution. We will have dual 10 Gb lines > > between the sites with a distance of 110 km (69 Miles). I am told the > > latency will be very low - as if the 2 servers were in the same data > > center. We are using top of the line, fast Unix Solaris servers. > > > > Does anyone have any input? Is anyone running a similar production setup > > with fast servers, many concurrent sessions and using HDR ? Any real > > production based feedback would be greatly appreciated. Also feel free to > > email me offline. > > > > Thank You, > > > > David > > in4mixdba@gmail.com > > > > --20cf3066791994ed9304c3ed8765 > > > > > > > > > > ******************************************************************************* > > Forum Note: Use "Reply" to post a response in the discussion forum. > > > > > > -- > Fernando Nunes > Portugal > > http://informix-technology.blogspot.com > My email works... but I don't check it frequently... > > --20cf303b40ef03b0d004c3ee12da > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --20cf306f74d866d02504c3ef01c7

I have used 11.10 with HDR and RSS to replicate from Canada to Malta. Had to add Riverbed WAN optimization due to the bandwidth into Malta and it worked great with zero lag. I have just setup 11.50.FC4 with 1 primary and 2 RSS servers. One located in next to the primary in the same data center, Memphis, TN and the second located at our data center in Elk Grove, CA. Works great, running 2.5 million SQL calls per minute. Both RSS servers have zero lag. IBM Informix Dynamic Server Version 11.50.FC4 -- On-Line -- Up 63 days 10:11:45 -- 68055168 Kbytes Local server type: Standard Index page logging status: Enabled Index page logging was enabled at: 2011/12/10 23:57:16 Number of RSS servers: 2 RSS Server information: RSS Srv RSS Srv Connection Next LPG to send Supports name status status (log id,page) Proxy Writes xxxxxxxxss Active Connected 165913,38975 N xxxxxxrss Active Connected 165913,38975 N