Disk usage for partition bootflash on fabric interconnect exceeded 70%

[robad]

Hi,

I'm getting this Minor fault on my primary FI :

Disk usage for partition bootflash on fabric interconnect A exceeded 70%

Wanted to know, what can I delete from it, except .bin files.

Is there some "garbage" files that I can delete from it ?

thanks in advance,

[Kirk J]

Greetings.

Have you removed all the unused Infrastructure, B series, and C series package firmware bundles?

The firmware is generally the largest consumer of space on /bootflash.

Make sure sure you have cleared out any old techsupport files.

Thanks,

Kirk...

[robad]

Hi Kirk,

1. TechSupport Files - On "Admin" tab under "TechSupport Files" I have nothing .

2. Unused Firmware Bundles - that's why I asked if there are another files except firmware packages, because, last time I had this error, I mistakelly deleted a version which was in used on the "HFP", and I had lot of errors .

Now I'm afraid and want to be sure that the version I'll delete is not in used on any place .

So how can I check it ? or, again the same Q, is there another files I can delete and it'll free some space, or only Deleting unused Firmware bundles will help ?

thanks a lot in advance,

[Kirk J]

For each of your host firmware package policies, you need to scroll through each one and note the Blade and Rack package it is referencing.  For any of the firmware packages that are not referenced in any of your host firmware policies, you can delete without getting errors.

You should also check your host firmware policies to see if they are even assigned to any blades/servers (show policy use), and potentially remove the host firmware package (or change the referenced firmware bundle to one you are retaining) that have no service profiles linked to them.

Occasionally you can have issues where one FI doesn't remove files on the 2ndary, or if there as ext3 file system issues on the /bootflash file system.

For those kind of issues you can do a nxos#show logging log

and look for ext3 related messages.

If you have FI /bootflash space utilization that is very different between the 2 FIs or see some ext3 errors in logging log output, then you'll want to open a TAC case.

Thanks,

Kirk...

[Ron Baduach]

Hi Kirk, thanks for the info

the packages I deleted didn't free much space

The bootflash usage on FI-A is 82% , and on FI-B is 39%

I can't open a TAC because the equipment is not under warranty anymore

do you have some more command I can check in order to get to a point where it'll be balanced ?

or even to only stop the fault message ?

thanks

[Kirk J]

If A and B are that far off then it sounds like there are some orphaned tech supports, backups, etc, or file system issues, that would take a dplug/root access to look through what are your larger files.

If you do manage to arrange a dplug session, you can run the following:

#find /bootflash -printf '%s %p\n'| sort -nr | head -20

which will give you a sorted, descending list of files by size in bootflash (top 20)

Thanks,

Kirk...

[ITM-Team · Accepted solution]

We had the same fault and could reduce the used space by deleting core files.

You can check via UCSM: Admin -> Faults, Events and Audit Log -> Core Files

As we had a lot of core files and problems displaying them in the Manager I deleted them with the command line. You need to log in to the corresponding FI:

FI-A # scope monitoring
FI-A /monitoring # scope sysdebug
FI-A /monitoring/sysdebug # clear cores
FI-A /monitoring/sysdebug # commit-buffer

After deleting the files the UCS needs a bit time to recognize the free space.

[robad]

Thanks ! It's solved my issue !

Kirk, I did't succeeded with running the command you gave

Thanks guys