Replacing 5800 HA ClusterXL with 9200s. Should I convert to ElasticXL?

[Perry_McGrew]

We have pair of 5800's in HA ClusterXL setup in our datacenter.   We have a pair of 9200's on the way.   All our CP GWs and Mgt are currently R82 JHF 39.

I have read a little about ElasticXL but not sure how to get there from our ClusterXL setup -- or if it really will see any benefit.   We are Healthcare and can't afford downtime -- been a long time since we upgraded from 4600 -> 5800. 

Typically, starting with the Passive GW, just do the basic setup on the new appliance, import the config, swap it in, establish SIC and  push policy (unchecking the "for gateway clusters...." option).  Once policy installed, fail over and repeat process to replace the other 5800.  

I have looked around and have not found any postings where anyone has done a conversion on live HA ClusterXL to ElasticXL.   If there is an outage needed, I probably will continue with ClusterXL.   

Any experiences, suggestions???

TIA - Perry

[the_rock]

I could be mistaken when I say this, but I dont believe you can "convert" clusterxl to elasticXL, would need to do fresh install, but maybe someone else can confirm.

Andy

[Chris_Atkinson · Accepted solution]

Currently it would be a parallel build and migrate (with downtime) to arrive at an ElasticXL cluster.

A conversion option is planned but not yet generally available, further details are not yet widely known.

See also: Working with ElasticXL Cluster https://sc1.checkpoint.com/documents/R82/WebAdminGuides/EN/CP_R82_ScalablePlatforms_AdminGuide/Conte...

[matanbe_chkpcp · Accepted solution]

Hello,

ElasticXL provides significant value in scaling, redundancy, day-to-day operations, monitoring, debugging, and more.

Regarding migration from CXL to EXL – we have a tool for that, which is currently in final testing but can be shared privately if needed. Please note that using this tool involves a short downtime, roughly equivalent to a reboot.

If you’d like more details, feel free to reach out here or contact me privately.

Regards,
Matan

[the_rock]

Good to know Matan! I may have a customer that would be interested.

Andy

[Perry_McGrew]

Matam,

Thanks for the reply and I have sent you a PM. Also thanks to @Chris_Atkinson and @the_rock for adding their input.  I will update the post if I do the conversion to let everyone know how it went, 

Perry

[the_rock]

Definitely let us know how it went if you do decide to use it.

Andy

[Perry_McGrew]

Matan,

The 9200's are in route.  Should be here next week.  Very interested to see how to migrate from our 5800 ClusterXL HA to the 9200 ElasticXL HA config.  

Regards,

Perry

[HeikoAnkenbrand]

Hi @matanbe_chkpcp,

I have a question regarding this topic.

When migrating from a ClusterXL to an ElasticXL cluster using the tool, I see the need to properly connect the systems (SYNC, MAGG1, ...). How will this be handled with the upcoming tool?

Does the tool also support VSX to VSNext?

[Perry_McGrew]

@matanbe_chkpcp 

I sent a private message a while back and have not heard back.   The 9200's are not here yet, but I'd like to know if this conversion tool that can convert our 5800 ClusterXL to 9200 ElasticXL is ready.

Regards,

Perry

[ZSEagle]

Hello,I need to migrate my devices that are currently in a cluster setup to an Elastic system. I have read your entries on the forum and understood that there may be a way to do this. Unfortunately, I am unable to contact you via private message. I would greatly appreciate it if you could get back to me.

[Perry_McGrew]

Unfortunately, I have not heard back either.   My 9200's are still in ClusterXL as well.  I have not seen any official tool or process to convert. Everyone here pretty much stated there is an outage involved.   I don't have a lab setup to test so I need to wait and see before committing to migrate to ElasticXL.  May be it is in R82.10 - which I am waiting for the 1st JHF to be released before upgrading from R82 JHF70

[the_rock]

Hey Perry,

Just curious, did you check to see if Professional services might be able to assist? I mean, Im positive there would be extra charge for it, but might be worth exploring that.

[Perry_McGrew]

I wish.... but we are not-for-profit Healthcare.   Rarely get the opportunity to use Professional Services.  Typically we do 99.99% of IT in house. 

Matan never responded back about that migration tool the development team was working on.   Our CP Account team has not heard anything either.   Really would like to go to ElasticXL -- but as a 24x7x365 medical facility w/ several sites, outages are frowned upon 😉 .

[the_rock]

I get it, but if outage is required and cant be avoided, not really sure what other option would be acceptable.

[Perry_McGrew]

I would just need to provide a pretty accurate indication of the length of any outage.   I'd map out each step in the process and need to have the fallback detailed as well. 

ClusterXL works...there is no real urgent need to move to ElasticXL.  The increase performance from 5800 to 9200 was significant in itself -- we ordered the 9200 pair with max RAM.  ElasticXL seems to be more efficient by using both GWs than ClusterXL's Active/Passive config.  

I'll likely look into it more later this year after moving up to R82.10 + JHF. By then, more CheckMates may have experiences with it that I can use as a guide.  

[the_rock]

That sounds reasonable.

[ZSEagle]

I have opened a support case regarding this issue. Once they provide feedback, I will share the updates with you as well.Based on my research so far, it appears that this migration may not be possible; however, I will keep you informed of any developments.

[the_rock]

Please keep us posted. Based on my logical assumption so far, sounds like it would be possible, but no way to do it without downtime.