Error attempting to upgrade 3550 with RME 3.5

[rickellis]

Hi,

Attempting to upgrade a 3550 switch using RME 3.5 (IDU9) and get the following error:

=============================

Error Info Code: '4175'

Message: 'Device Error: Flash copy to/from operation failed.'

Action/Reason:

'Probable Cause: The source file size and the destination file size do notmatch. Disk space on tftp root directory may not be sufficient.

Recommended Action: Verify disk space in the TFTP directory and retry the operation.

===============================

I have lots of room in the TFTP root on Ciscoworks (45GB) and I know that the job is working because it deletes the old IOS completely it just won't transfer the image over. I have plenty of room on the 3550 (15mb) for the 4mb incoming IOS image.The switch runs 12.1.20 EMI attempting to upgrade to 12.1.22 EMI.If I perform the same operation manually it works fine. Any ideas?

thanks,

Rick

[rmushtaq]

First turn on the SWIM debug under RME->Administration->Software Management->Edit Preferences. Check the box for 'Turn debugging on'. Then click Finish. Then run the SWIM job again. Once it fails, goto RME->Software Management->Job Management->Mail or Copy Log File. Select the log files and save them to a location. Then look at these log files which should tell exactly where/why the SWIM job is failing. Make sure that your turn off the Debug flag once done.

[rickellis]

Hi,

I've enabled the debug steps and I see that the TFTP transfer starts and times out as follows:

====================

Loading px_sw_12141_7092 from 10.254.64.2 (via Vlan64): !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!..O. [timed out]

%Error reading tftp://10.254.64.2/px_sw_12141_7092 (Timed out)

====================

I've set the exec timeout on vty as:

line vty 0 4

exec-timeout 0 0

line vty 5 15

exec-timeout 0 0

I've actually put the switch in question on the same segment as the Ciscoworks server to eliminate any possible issues like ACL's or WAN link speeds.

Any help appreciated.

thanks,

Rick

[minie]

Is the "cwcs tftp service" running? Try do "copy tftp flash" from the device thru CLI, see if tftp is successful.

[minie]

I had a second look on your post and realize you have already done what I said. I have a few more thoughts as this sounds too odd.

- image integrity

Check the file size of "px_sw_12141_7092" under tftpboo t directory. It might be a corrupted image. Chances are image got tftped over, but failed on final checksum

- RME delete px_sw_12141_7092 file before tftp is done.

So you need to watch tftpboot directory frequently during upgrade process

[rickellis]

I checked the image under tftpbootdirectory. The folder is empty aside from a readme stating to keep the directory. I tried another upgrade and saw that the entire IOS image gets copied into the tftpbootdirectory (the image size looks right) and simultaneously did a dir listing on the 3550 being upgraded. I saw about 1/2 mb of the image on the switch (so I know it's trying), then it just stopped. I re-checked the tftpbootdirectory and the IOS image was gone (presumably this happens when the transfer fails or is a normal occurence with upgrades I'm not sure). Overall the TFTP is working it just times out for whatever reason at about 1/2 mb. I have 15mb free flash so that shouldn't be an issue.