topic Re: Brocade flash monitor marginal in HPE EVA Storage

Brocade flash monitor marginal

David ASF — Sat, 09 May 2009 06:21:36 GMT

Hello
after upgrade to fabric OS 6.2.0d, my switch report a marginal state for flash monitor. In the log, I can find a line :
warning : res flash 0, is above high boundary (high = 85, low=0). Current value is 87 %

any idea ?

thank's

Re: Brocade flash monitor marginal

IBaltay — Sat, 09 May 2009 07:21:59 GMT

Hi,
look at e.g. the fabos 6.1.1 admin guide, from page 49, switch status policy treshold values section:
http://h20000.www2.hp.com/bc/docs/support/SupportManual/c01613852/c01613852.pdf?jumpid=reg_R1002_USEN

and also FABOS 6.2.0 message reference guide:
http://h20000.www2.hp.com/bc/docs/support/SupportManual/c01711366/c01711366.pdf?jumpid=reg_R1002_USEN

page 287: (fw-1400 - 1403)
yours seems to be fw-1402:
you can find a recommended actions there too

Re: Brocade flash monitor marginal

David ASF — Sat, 09 May 2009 10:00:14 GMT

Hi
thank's for your help. I've try using supportsave command. It uploaded 1,3Mb of log files to my ftp server but I still have the alert after. I have disable the alert using the supportpolicyset command but do you think it's normal to have 87% of the flash memory use by the system ?

David

Re: Brocade flash monitor marginal

IBaltay — Sat, 09 May 2009 11:06:41 GMT

When you requested a firmware download during the firmware upgrade, the system first checked the file size being downloaded. If the compact flash did not have enough space, the download/upgrade did not occur. Now its size is a bit bigger then the default size treshold of 85%.

here are the 6.2.0d release notes for more details on e.g. supported SAN switches, etc.
http://h20000.www2.hp.com/bc/docs/support/SupportManual/c01709493/c01709493.pdf?jumpid=reg_R1002_USEN

the very similar problem to yours is listed there on page 41 - (fixed in the 6.2.0b? :-))

Re: Brocade flash monitor marginal

IBaltay — Sat, 09 May 2009 11:28:57 GMT

in other words if the size is 87% and the interactive mode does not store the supportsave in the switch root directory (according to the RN fixed with 6.2.0b), you seem to live with the 13% percent of free flash, without notifying it...

Re: Brocade flash monitor marginal

Bill Rothanburg — Mon, 11 May 2009 13:52:18 GMT

You might have some old core files that can be deleted. Look in the supportshow file from your supportsave for the line "find /core_files /mnt/core_files -type f". If you see files begining with core.pd or misc.pd, then these could be the cause of the flash being so full.

The "supportSave -R" command can be used to remove them.

Re: Brocade flash monitor marginal

David ASF — Mon, 11 May 2009 19:00:11 GMT

Hi
supportsave -R command return No core or FFDC data files found!
SupportSave completed

David

Re: Brocade flash monitor marginal

Tim Chilton — Tue, 12 May 2009 15:53:54 GMT

We have had this problem as well after we upgraded to v6.2.0.a in February. We have been instructed to run supportsave -R, and that didn't work for us either. We have also been told to use the cleanup command, which is pretty much undocumented and must be run as root. Neither has worked for us. HP tier 3 is working with Brocade for a fix, but it's starting to look like we're going to be replacing switches. If there is any change from my side, I'll post.

Re: Brocade flash monitor marginal

RenatoMartini — Thu, 13 Oct 2016 02:51:05 GMT

...use the command 'cleanup' to clean obsolete files. So, may decrease the used space in the switch FOS file systems.