High snmpMaster Memory Usage x450
Options
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Report Inappropriate Content
‎12-02-2014 04:54 PM
Recently we have had some trouble with some of our older extremes rebooting. Looking at debug logs it appears to be pretty clearly a depletion of system meomory. This has forced us to take a closer look at the memory of these devices and one thing that we came across was that the SNMP process appears to be utilizing ~27% of the available system memory:
Slot-1 Switch_.11 # sh memory process "snmpMaster"
System Memory Information
-------------------------
Slot-1 Total DRAM (KB): 262144
Slot-1 System (KB): 7480
Slot-1 User (KB): 224536
Slot-1 Free (KB): 30128
Memory Utilization Statistics
-----------------------------
Card Slot Process Name Memory (KB)
---------------------------------------
Slot-1 1 snmpMaster 72024
Slot-1 Switch_.16 # sh memory process "snmpSubagent"
System Memory Information
-------------------------
Slot-1 Total DRAM (KB): 262144
Slot-1 System (KB): 7480
Slot-1 User (KB): 224544
Slot-1 Free (KB): 30120
Memory Utilization Statistics
-----------------------------
Card Slot Process Name Memory (KB)
---------------------------------------
Slot-1 1 snmpSubagent 2720
This seems to be happening even when our monitoring system is in pause mode for that device. Is this normal, or is there any way that we can reduce the memory usage of this process? I have checked release notes and it appears there was at some point memory issues with snmp process, but they should have been fixed with 12.6.2-patch1-6
We are on:
Image Selected: primary
Image Booted: primary
Primary ver: 12.6.2.10
Secondary ver: 12.0.4.5
and
Slot: Slot-2 * Slot-1
------------------------ ------------------------
Current State: MASTER BACKUP (In Sync)
Image Selected: primary primary
Image Booted: primary primary
Primary ver: 12.5.1.6 12.5.1.6
Secondary ver: 12.0.4.5 12.0.4.5
and
Slot: Slot-1 * No Backup
------------------------ ------------------------
Current State: MASTER
Image Selected: primary
Image Booted: primary
Primary ver: 12.6.2.10
Secondary ver: 12.0.4.5
all of these switches have had this issue of rebooting due to memory depletion.
Slot-1 Switch_.11 # sh memory process "snmpMaster"
System Memory Information
-------------------------
Slot-1 Total DRAM (KB): 262144
Slot-1 System (KB): 7480
Slot-1 User (KB): 224536
Slot-1 Free (KB): 30128
Memory Utilization Statistics
-----------------------------
Card Slot Process Name Memory (KB)
---------------------------------------
Slot-1 1 snmpMaster 72024
Slot-1 Switch_.16 # sh memory process "snmpSubagent"
System Memory Information
-------------------------
Slot-1 Total DRAM (KB): 262144
Slot-1 System (KB): 7480
Slot-1 User (KB): 224544
Slot-1 Free (KB): 30120
Memory Utilization Statistics
-----------------------------
Card Slot Process Name Memory (KB)
---------------------------------------
Slot-1 1 snmpSubagent 2720
This seems to be happening even when our monitoring system is in pause mode for that device. Is this normal, or is there any way that we can reduce the memory usage of this process? I have checked release notes and it appears there was at some point memory issues with snmp process, but they should have been fixed with 12.6.2-patch1-6
We are on:
Image Selected: primary
Image Booted: primary
Primary ver: 12.6.2.10
Secondary ver: 12.0.4.5
and
Slot: Slot-2 * Slot-1
------------------------ ------------------------
Current State: MASTER BACKUP (In Sync)
Image Selected: primary primary
Image Booted: primary primary
Primary ver: 12.5.1.6 12.5.1.6
Secondary ver: 12.0.4.5 12.0.4.5
and
Slot: Slot-1 * No Backup
------------------------ ------------------------
Current State: MASTER
Image Selected: primary
Image Booted: primary
Primary ver: 12.6.2.10
Secondary ver: 12.0.4.5
all of these switches have had this issue of rebooting due to memory depletion.
1 REPLY 1
Options
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Report Inappropriate Content
‎12-02-2014 08:56 PM
Hi Evan,
At first glance, this seems to be related to PD4-2345650880 - "The snmpMaster process experiences memory depletion of approximately 4096 bytes per SNMP polling."
Notes in the PD indicate that restarting the SNMP process can recover memory, as does a reboot. To restart the snmpMaster process, use this command:
restart process "snmpMaster" You've mentioned that you have EXOS 12.6.2.10 on two of your switches, so I'm mentioning the PD that is for that release. It is fixed in EXOS v12.6.2.10-patch1-2 (and greater)
For 12.5.1.6, you'll need to upgrade to at least EXOS v12.5.4.5-patch1-7 (per PD4-2198529731).
However, all versions of EXOS 12.x are at the end of support. I would recommend that you upgrade to the latest patch version of EXOS 15.3.x, as that is the last release version supporting the X450a/e series. It will also include the aforementioned SNMP fixes.
Again, I'm not positive this is your issue, but it does seem likely.
I hope this helps, and sorry to hear of your troubles.
-Drew
At first glance, this seems to be related to PD4-2345650880 - "The snmpMaster process experiences memory depletion of approximately 4096 bytes per SNMP polling."
Notes in the PD indicate that restarting the SNMP process can recover memory, as does a reboot. To restart the snmpMaster process, use this command:
restart process "snmpMaster" You've mentioned that you have EXOS 12.6.2.10 on two of your switches, so I'm mentioning the PD that is for that release. It is fixed in EXOS v12.6.2.10-patch1-2 (and greater)
For 12.5.1.6, you'll need to upgrade to at least EXOS v12.5.4.5-patch1-7 (per PD4-2198529731).
However, all versions of EXOS 12.x are at the end of support. I would recommend that you upgrade to the latest patch version of EXOS 15.3.x, as that is the last release version supporting the X450a/e series. It will also include the aforementioned SNMP fixes.
Again, I'm not positive this is your issue, but it does seem likely.
I hope this helps, and sorry to hear of your troubles.
-Drew
