Cisco switch

[Mlex1]

Hello everyone i see logs on the switch.

%SYS-2-MALLOCFAIL: Memory allocation of 395648 bytes failed from 0x1AE504, alignment

%Error on initialize VLAN database 1: VTP feature not yet initialized[OK]

googled this issue, found memory is low, how to solv the problem?

show memory statistics
Head Total(b) Used(b) Free(b) Lowest(b) Largest(b)Processor 2705AEC 22233908 21542640 691268 111792 519596 I/O 2C00000 4194304 2575564 1618740 1110476 1613888Driver te 1880000 1048576 44 1048532 1048532 1048532

[Mark Elsen]

  - @Mlex1           Memory leaks can be caused by software bugs ; use the latest advisory release on the platform

  M.

[Mlex1]

i have the last,

Cisco IOS Software, C2960 Software (C2960-LANBASEK9-M), Version 15.0(2)SE11, RELEASE SOFTWARE (fc3)

[Mark Elsen]

  - @Mlex1                     Then  report the issue to TAC
                                      ( Added 🙂  or have a look at : https://www.cisco.com/c/en/us/support/docs/ios-nx-os-software/ios-xe-17/225227-troubleshoot-memory-issues-in-ios-xe.html

  M.

[Joseph W. Doherty]

googled this issue, found memory is low, how to solv the problem?

Assuming you're not bumping into a bug, either adjust config to use less memory and/or (if possible) add RAM to device.

BTW, sometimes low memory causes issues not because you've actually exhausted it, but because it's fragmented to an extend a large enough free memory block doesn't exist.  However, mitigation is the same, reduce memory need and/or add RAM.

[Giuseppe Larosa]

Hello @Mlex1 ,

your switch runs classic IOS not IOS XE but likely the link provided by @Mark Elsen can be helpful

memory leakage increases with system uptime. So if memory usage is very high you can start by reloading the switch.

Once the switch comes up again you likely will be able to perform some actions on the switch.

If this kind of error comes out at boot or with few minutes uptime there is likely some RAM chip faulty and I'm afraid in this case the switch has to be dismissed.

Hope to help

Giuseppe

[Joseph W. Doherty]

---

@Giuseppe Larosa wrote:

memory leakage increases with system uptime. So if memory usage is very high you can start by reloading the switch.

---

@Giuseppe Larosa is 100% correct that the common immediate "fix" for many memory resource allocation issues is, unfortunately, a reload.

Just want to further add, again, not all memory allocation issues are due to memory leaks, i.e. bugs.  Some features, do "naturally", consume more and more memory over time and/or they may dynamically expand and contract over time.  Those features that dynamically expand and contract might also be very variable in how much memory they consume at any particular point in time.  I.e. it's possible you might go months and months, and one day, you have a memory issue, that after reloading, possibly, you'll never see again, or see it very infrequently.

Issues due to fragmentation, sometimes, can be precluded by reloading the device on some fixed schedule, such as monthly, weekly or even daily.

BTW, trying to find more info on your particular error message, memory "alignment" errors appear often are due to memory fragmentation rather than lack of memory (although low free memory often helps create fragmentation).

Also again, for non-bug memory consumers, adjusting feature usage may preclude the issue.

[Mlex1]

agree with both @Giuseppe Larosa  and @Giuseppe Larosa two time i already reload after a few minut i can't connect to the switch

[Joseph W. Doherty]

two time i already reload after a few minut i can't connect to the switch

Oh my, that's not a good sign.

What you may need to do disconnect the switch from the network and reload it and see if it appears stable.  Then try going to an absolute minimum config that supports basic network usage and reconnect it and see if it's stable. If so, you can try layering on additional features while watching for any instability.

You might also consider trying using c2960-lanbasek9-mz.122-55.SE12.

[aleabrahao]

I've seen similar cases, and it usually happens when you're using many features, such as SNMP polling, logging buffers, ACLs, spanning-tree instances, QoS policies, etc.

One thing you can try is disabling or reducing the logging buffer.

Run the commands: `show processes memory sorted` and `show memory allocating-process totals`

and look for processes growing continuously.

Since this is a very old switch and no longer has updates/bug fixes and vulnerability patches, I also suggest that you replace it.