SW_DEDUPE_INLINE_NON_STANDARD events logged after enabling inline deduplication

Summary: After enabling inline-deduplication, various SW_DEDUPE_INLINE_DEGRADED 400180004 & 400,180,005 events may occur.

This article applies to This article does not apply to This article is not tied to any specific product. Not all product versions are identified in this article.
Check out other resources

Symptoms


One or more of the following event messages may be generated after enabling inline deduplication. Corresponding Email events may also be sent from the cluster.
  
58.125786  01/09 15:20 W    49   215295         Index layout is non standard   
58.125785  01/09 15:20 W    49   215296         Inline dedupe running degraded with smaller index      
 
SW_DEDUPE_INLINE_DEGRADED 400180004 - Warning. 
The index has been allocated with less than the requested size due to the load on the node. 
Inline dedupe is running but with a reduced window into the set of blocks that can be deduplicated. 
Disabling inline dedupe on the node and trying the allocation again may succeed, but is unlikely to work unless the load on the node has dropped significantly. 
The only reliable way to fix this is to reboot the node so the kernel can easily allocate the index at boot time.
SW_DEDUPE_INLINE_NON_STANDARD 400180005 - Warning. 
The index has been allocated at the requested size and inline dedupe is running but the kernel could not allocate the best layout because of the existing load on the node. 
The top level table is larger than it could be, taking up more kernel pages. 
Disabling inline dedupe on the node and trying the allocation again may succeed, but is unlikely to work unless the load on the node has dropped significantly. 
The only reliable way to fix this is to reboot the node so the kernel can easily allocate the index at boot time.

Cause

This can occur if inline deduplication is enabled several days or weeks after a node has already booted up and has a large uptime. Memory in the nodes may have already been allocated for other cluster functions so Inline Deduplication may not be working at optimal efficiency when first enabled. These warnings indicate that inline deduplication is still running, but at less efficiency.

Resolution

In many situations, this issue self-corrects itself as each node reallocates memory to accommodate inline deduplication. This may take several hours depending on cluster size and workload. The cluster stops logging the events once the issue is fully self-corrected.
 

In versions of OneFS prior to 9.5.0.2, these events were logged as Warning Events. Later versions of OneFS with newer Roll Up Patches log these as Informational Events. In either case, no actions are typically necessary. 
 

If a node is unable to self-correct due to ongoing workload on the cluster, a reboot of the affected nodes may be necessary. Rebooting nodes forces reallocation of memory at boot time to accommodate inline deduplication.
 

If these events continue, reach out to Dell Technical Support for assistance. 

Affected Products

PowerScale OneFS

Products

Isilon
Article Properties
Article Number: 000228179
Article Type: Solution
Last Modified: 11 شعبان 1447
Version:  3
Find answers to your questions from other Dell users
Support Services
Check if your device is covered by Support Services.