This post is more than 5 years old
124 Posts
0
1011
Active teir is filling up
Hi My active tier seems to be filling up. I see movement to the archive tier but does not seem high enough.
This is the settings at moment.
What else can i check please?>
This post is more than 5 years old
124 Posts
0
1011
Hi My active tier seems to be filling up. I see movement to the archive tier but does not seem high enough.
This is the settings at moment.
What else can i check please?>
Top
ble1
14.3K Posts
0
March 27th, 2015 02:00
As I said, I do not have experience with archive tier, but top of my head I would say there is always some kind of trade-off. Otherwise one would always use just one approach. I suspect that too often movement does have some impact on resources so too often might not be good. 1 day sounds as if someone wants to keep data on archive tier instead of active one so I have to ask - do you run backup or archive there? If you use this just to free up active tier, think about following:
- more load on DD with data movement
- slower restore (I'm not sure, but if restore brings data back to active tier, then I can see how one can easily shoot himself/herself in the foot with 1 day policy)
So, I think you need to step out of panic mode and get control of what is going on and why. For that, you need support surely and I think that support goes beyond chit-chat on these forums. I would advise to contact support and get things investigated together. Since it is Friday, you do not have much time I guess
ble1
14.3K Posts
0
March 27th, 2015 01:00
Buy more storage
Seriously, if sizing is not good (eg. filling up or living on the edge), then you must address real needs. I'm not sure if that involves (in your case) future growth, but it doesn't sound like it does. There are 10 phases of cleanup; you can see % of it for each when running filesys clean status. You need to survive until phase 9 which is when space gets reclaimed back. Meanwhile, check your policies and usage and based on that calculate real needs. You wish to have situation where primary tier - ideally - never reaches 80%. But that would be ideally so I guess more practically that would be 90%, but that is already dancing on the edge.
boerbokrib
124 Posts
0
March 27th, 2015 02:00
Hi
This has only started happening in last few days my backups are not that big so don't understand why it is happening.
Can you explain this to me please.
When data movement starts, all files that have not been modified for the specified threshold number of days are moved from the active tier to the retention tier.so is a higher number of days or lower number of days better here?
I know . What happens when active tier fills up. does the data get moved to archive tier?
ble1
14.3K Posts
0
March 27th, 2015 02:00
I guess answer depends on how is your env manager (is it application based backup which is integrated or not). I do no have experience with archive tier, but I suspect that there are different indices for active and archive tier (and while I do not know this, I would imagine transfer from active to archive tier like replication from one DD box to another). Anyway, if your active tier started to fill up only recently, there are couple of things to check:
- is there any new data which might not compress/de-dupe well?
- is there anything on backup (application) side which might not mark data to be removed from DD?
- do you see difference in cleanable GiB column in df output on DD compared to previous times
Last one is kind of thing I like to check normally as it provides trend info on usage and cleaning cycle. If I see difference, I know something must have happened. I recently had similar issue (without archive tier) where my active tier started to fill up due to backup application marking volumes with specific flag which instructed backup application cleanup procedure to skip those volumes; as consequence data there which has expired was not marked expired and DD would not consider it for its own cleanup.
In your case you have extra layer - archive tier. It is not clear to me if what you have is caused by:
a) active tier filling up more quickly due to more data
b) active tier filling up more quickly due to slower (less) transfer to archive tier
c) both
d) none of it
I assume, one would need to check past autosupport reports to get an idea on how this was trending in the past and if only started to deviate more recently then find the date when it started. Then you focus on logs and event around that data until you don't find root cause. This might be done more quickly with assistance of support who might already know very well what to look or where to look (and if your autosupport is sent to EMC, most likely they can easily get trending information).
boerbokrib
124 Posts
0
March 27th, 2015 02:00
Can you explain this to me please.
When data movement starts, all files that have not been modified for the specified threshold number of days are moved from the active tier to the retention tier.so is a higher number of days or lower number of days better here?
boerbokrib
124 Posts
0
March 27th, 2015 04:00
Hi It appears that the supplier of the datadomain has setup a replication of vtl and that they say is what is filling up the space. I would love to log call with support but all has to go through the supplier.
I see that some space has freed up on the active tier when it hit stage 9 you the boss.
I will set it to 1 day for moment and run disk movement until they have completed the replication.