This post is more than 5 years old
261 Posts
0
3265
source centera not having enough space
I have two centera Gen4 boxes : Primary and remote with bidirectional replication.
Recently we had centrastar upgraded to-4.1.0,but soon started having replications stopped because the source cluster was apparently not having enough space.
The source Centera is running out of capacity. Only 279 GB of space is available.and EMC is insisting on adding more capacity to resolve the parking issue.
I also a the garbage collector to reclaim a sweep space of 121 GB.
I need to convince the customer to opt for adding more nodes to the source cluster . What should be my convincing statement ?
At this point : Replication is paused and there is parking overflow.
What 're the implications of not heeding to the above issue and letting the replication be in puased state with paking .
Thanks in advance!
Anonymous
5 Practitioner
5 Practitioner
•
274.2K Posts
0
May 10th, 2011 08:00
Hi Deeppat
There are several major implications:
The removal of new data protection provided by the replica (depending on protection mode Centera is set to - one or two copies are not created),
The application host or hosts will not be able to write any new data to this source Centera, causing application host filling up its own cache, effecting performance, etc.
The host reads performance will also be degraded.
Internal Centera self healings are taking longer….
As part of protection against such event, the designers of centera implemented a special parameter setting “suicide prevention” which is used to put centera into read only mode when available capacity is less than 5%.
I hope this helps.
Anonymous
5 Practitioner
5 Practitioner
•
274.2K Posts
1
July 27th, 2011 02:00
deeppat, you could have a check for the FreeObjectCount first, to see if space is available but no ObjectCount remains, this situation of low available FreeObjectCount would appear the same problem as NoCapacity.
if that's not the problem, which protection mode is in this Replicator pair? if CPP, what's the size of the parameter configured for the parity write size? Uploading files less than above parameter size would largely waste the capacity in the cluster.
hope this helpful
holgerjakob_c0722c
337 Posts
1
August 11th, 2011 22:00
Hi
A quick fix might also be to lower the supported object count if you are capacity and not object bound.
If you want me to check please send the output of a sho capacity tot and sho capacity ava to holger.jakob@informatio.ch.
Our managed services would have told the customer at least 18 months before that he would be hitting either a capacity or object count limit. That the Centera does require a minimal amount of monitoring should also be communicated as well as the fact that adding nodes will be necessary for his applications to continue to be able to archive.
Running a Garbage Collection might also free up some space. We have seen many many systems where GCii was not running.
Just let me know if you need assistance, Holger
deeppat
261 Posts
0
October 4th, 2011 12:00
thanks all for the response.The Customer has agreed on adding more nodes to the source cluster.
Can any one please point me at the right DOCUMENT which has information pertaining to it?
I need to know what all things are needed while adding modes to the cluster. Aprreciate if some one can summarize this a little.
Thanks in Advance.
GBcentera
46 Posts
1
October 5th, 2011 04:00
To add nodes to a Cluster I would recommend reviewing the procedure for adding nodes. The procedure can be generated from the Centera Procedure Generator.
Open the application and select the path CS Procedure > Centera > Add Hardware ? Add Gen4LP – 4/8TB 2-Node Increment.
Also please follow the CCA process and submit and CCRF for review.
deeppat
261 Posts
0
October 14th, 2011 13:00
Hello ,
I have a follow up question:
I ran the IC for CICM migration for the cube i.e cube 1 to cube 3 and cube 4 . After the completion of the migration (basically, what my task is ; Migrating cube1 and cube 2 to cube 3 and Cube4).
There were leftovers around 4 and 5 clips not migrated after the IC was run. After that , I had to raise a case with support.They ran Centera exerciser to move those stuck clips and blobs and confirmed the stuck blobs are moved. Now , when i run show migration detail , i still see those 4 leftovers and 5 clips .
I am not sure why this is showing up?? Am i okie to proceed with accepting the migration??
Please suggest !!