This post is more than 5 years old
2 Intern
•
522 Posts
0
2157
July 2nd, 2015 05:00
XtremIO Integration with RecoverPoint Technote Questions
Hi,
I am cross-posting this from the XtremIO ECN site since there have been no responses provided:
Hello,
After reading the integration technote with XtremIO and RecoverPoint, I have some clarification questions I am seeking some confimraiton/answers to please:
XtremIO-to-XtremIO Replication is assumed here:
1. When XtremIO is the source array, there looks to be about 2 snaps at any given time and an auto-created XtremIO CG for the volume being protected with RP – are these snaps/CGs identifiable in the XtremIO GUI or under the covers? How would a user query these?
2. When the XtremIO is the target array, there can be a max of 500 snaps created – are these XtremIO snapshots that are created identifiable in the XtremIO GUI or under the covers? How would a user query these?
a. Is there an XtremIO CG auto-created on the target XtremIO array with this solution for any failover tasks that might occur, or just the source XtremIO array?
3. The technotes have a step for the user to create an XtremIO CG and put in all of the volumes that RP will replicate:
“On the XtremIO management server, create an consistency group that contains all
XtremIO volumes involved in RecoverPoint replication.”
>>> Why is this needed if RP will automatically create the XtremIO CG’s per RP CG?
>>> How is the limit of 256 volumes per XtremIO CG handled in this setup? Just create multiple XtremIO CG’s to handle 256 volumes each?
4. There is a snippet on how image access is handled in the technote:
“The array actually presents a copy of the snapshot. The original snapshot is not affected. Replication and distribution continue, unaffected by the image access.”
>>> Does this mean that for the snapshot created on the target array that the user selects in RP's journal, another XtremIO snapshot is created for host presentation since it presents a copy of the snapshot? If so, do these “copies” get accounted in the overall snapshot limits of the array? This is a little different verbiage than earlier in the document which just says it access the actual snapshot so I'm trying to see if it really creates more or not here.
Thanks!
Idan
675 Posts
0
July 2nd, 2015 06:00
Hi Keith,
1) The CG is identifiable but the snaps and snap sets are only visible through the user rp_user which is a special user reserved for RP operations. The XMS is registered in the RP using that user.
2) 500 snaps per volume is the maximum, less can be configured on a per RP copy policy basis. each PiT on RP is a snap on the target XtremIO array. The user can see these snaps through RP's journal tab under the target copy. From the XtremIO CLI, these snap volumes will show up as internal volumes as they are visible only through rp_user.
a) All aspects of snapshot management in XtremIO are automatically managed by RP, nothing should be created automatically. This relates to both source and target arrays.
3) I'll get the TN fixed, there's no need to create CG but there's a need to register the RPAs and mask volumes (production, journals, repository, etc.) to them. With RP 4.1.2 and XtremIO 4.0, it's supported to have a max of 64 volumes per RP CG
4) So, in image access with XtremIO at the target, the PiT data isn't in journal. The Journal contains only metadata. the PiT is actually a snap on the target XtremIO array. Every time a snap is created in XtremIO, another volume (besides the derivative snap) is created, referencing the snap. That volume has an external SCSI personality, which means it can be exposed to a host.
Hope that helps,
Idan Kentor
Corporate Systems Engineering - RecoverPoint and VPLEX
echolaughmk
2 Intern
•
522 Posts
0
July 2nd, 2015 06:00
Hi Idan,
Thanks for the answers! You know me...had some follow-ups
.
#1/#2 - sounds like only the CG's will be seen by regular users (admin) and the actual internal volumes will not be noticeable by any user other than rp_user, which no one will use to log into the GUI/CLI. From a target perspective, given the Recoverpoint snapshots are related 1:1 to an actual XtremIO snapshot on the array, is there any command or reference a user could look at to quickly sum up the number of XtremIO snapshots on the array to see where they stand with regards to the total limit of 8192? I'm making the assumption that any user other than rp_user can't see these internal volumes with the native CLI/GUI.
a. Thinking maximum case scenario, if someone keeps 500 snaps of multiple RP CG's, since there is no ability to purge these manually, I'm just curious if they can be summed up in a way where a user could see what is consumed and modify the CG policy as needed in the case that they need to create manual XtremIO snapshots for their environment but get an error doing so because they would exceed the 8192 limit?
#3 - Thanks for that clarification, that helps. I didn't see that in the docs I read. That is good to know given parallel bookmarking isn't supported currently when using XtremIO volumes so we will have to get creative when a server/application has more than 64 LUNs that comprise it.
#4 - That sounds like it is just using the default snap mechanic of XtremIO so I am good there.
Thanks again for all the clarification and answers.
-Keith
echolaughmk
2 Intern
•
522 Posts
0
July 7th, 2015 06:00
I received some answers to the original post and wanted to cross-reference it here for completeness. I also included my follow-up questions to that post as well in case those answers happen to be known:
Re: RecoverPoint with XtremIO Technote Questions
#3 sounds like it might be 256 based on Tamir's response, so that technote might not need to be updated if that number is correct.
Thanks!
Idan
675 Posts
1
July 7th, 2015 07:00
Hi Keith,
It's 64 volumes per RP CG in case of XtremIO.
You can see the snaps (user snaps and the snaps that RP created) via the GUI and CLI command show-volumes. The user would get an error if going over 512 snaps per volume (500 in which RP can create) or a total of volume in a system.
Hope that helps,
Idan
echolaughmk
2 Intern
•
522 Posts
0
July 7th, 2015 08:00
Thanks Idan.....appreciate the clarification!
This is very helpful!
-Keith