This post is more than 5 years old
1 Rookie
•
23 Posts
0
6824
March 26th, 2015 22:00
RecoverPoint for VM - image access log
Hi Guys,
I have a question in regards to the image access log capacity when performing a “test copy” to test a copy in recoverpoint for VM.
If the test was long enough so the image access log fills, that is likely to stop all writes to the copy that is being accessed and therefor kill the test… is there an option to do a “test copy” and at the same time not having the source sending data to the destination so the image access log won’t fill to make sure the test wont be impacted?
No Events found!
Idan
675 Posts
0
March 26th, 2015 23:00
Hi there,
Let me start off by saying that the image access log is a limited portion (by default 20%) of the journal, it's allocated for writes on the target side during logged image access. There's an option to undo writes during image access if that space runs out.
During logged image access, data is sent to the journal volumes and distribution to the replica volumes is paused. Since a certain image is accessed, replication will not be disrupted as long as this accessed image isn't the image due to be overwritten (the oldest image) . In other words, the duration of image access w/o disruption of replication is a function of the Protection window.
Going back to your question though, you can pause replication which will allow you to test for indefinite time, assuming the image access log won't become full. if there's such a concern, you can add journal VMDKs or increase the percentage of the image access log using REST. Doing the latter will result in a journal history loss so if there's a need I would recommend to move to Direct image access.
Hope that helps,
Idan
mailto:idan.kentor@emc.com
maxsven
1 Rookie
•
23 Posts
0
March 27th, 2015 06:00
Hi Idan,
Thanks for the response.. I have the following questions:
when you said "Going back to your question though, you can pause replication which will allow you to test for indefinite time, assuming the image access log won't become full. " ------ Why would image access log ever become full if replication is paused?
Also, there is not option to pasue replication via the Recoverpoint for VM vSphere web client interface, is that going to be a CLI?
something like:
pause_transfer group=group1
then to resume:
start_transfer group=group1
Also, when you said:
"if there's a need I would recommend to move to Direct image access"
How do you do that? again no option in the interface to do that.
Idan
675 Posts
0
March 29th, 2015 07:00
Hi,
1) Image access log can become full when writes to the replica volume exceeds the capacity allocated for image access log (by default, 20% of the overall target journal capacity)
2) Use the pause_transfer/start_transfers CLI commands to pause replication for a specific group.
3) To enable direct image access, perform "Test a copy" via the plugin. Verify that the selected image is the relevant one (move to a different PiT if required) and then perform the set_image_access_mode CLI command to change to Direct image access. Afterwards, disable image access using the plugin. Not that Direct image access entails journal history loss (that's why it's critical to make sure that select image is the right one). When direct image access is disabled, only the changes b/w production and the replica will be transferred (AKA short sync).
Hope that helps,
Idan
ElBery
1 Message
0
January 24th, 2018 07:00
Hi, great post.
How can I increase the image acess log percentage via cli?
forshr
2 Intern
•
1.1K Posts
0
January 26th, 2018 00:00
By using the config_copy_policy command.
Regards,
Rich