Corrupt IOS

Answered Question
May 17th, 2010

I have a 3750G switch that has repeatedly developed a corrupt IOS file. It has happened 5 or 6 times this year after which I would reload the IOS and then everything will work fine for a month or so, then it happens again. Can anyone tell me if they have experienced this problem before and do you think it is a hardware problem?

Thanks

I have this problem too.
0 votes
Correct Answer by Leo Laohoo about 6 years 6 months ago

You've unfortunately hit a hole-in-one in terms of IOS bug and it's a very bad one too.  IOS upgrade is the only way to fix it.

Message was edited by: leolaohoo

Correct Answer by Leo Laohoo about 6 years 6 months ago

Giuseppe's right (+5).  Everytime the switch crashes it creates a file or subdirectory called crashinfo.  If the flash isn't cleaned out and it keeps crashing the switch will go to a state called crash-n-boot loop.  The switch will crash, creates a crashinfo file/subdirectory.  But since there's no more room at the "inn" or flash the switch will crash until someone will have to intervene.  Initially it will look like the IOS is corrupted but it isn't.

To delete subdirectories, use the command "delete /f /r flash:"

Now even if you've cleaned out the flash and you still get this error message about the "mzip files" then as Giuseppe has said, IOS upgrade is the way to go.  You have two avenues of doing this:  you can upgrade the IOS and boot the switch OR you can pump the IOS (if you have enough flash space) and wait for the next crash.

Hope this helps and don't forget to rate useful posts.

Correct Answer by Giuseppe Larosa about 6 years 6 months ago

Hello Mark,

you should consider an IOS upgrade.

if I've understood you are meaning that the switch crashes and each time a crash file is written on flash

the IOS image is loaded in memory at startup.

Do you check the file on flash with verify comand and you find that is corrupted?

if so you should think of finding a way to replace the internal flash

Hope to help

Giuseppe

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (4 ratings)
Loading.
Correct Answer
Giuseppe Larosa Mon, 05/17/2010 - 09:26

Hello Mark,

you should consider an IOS upgrade.

if I've understood you are meaning that the switch crashes and each time a crash file is written on flash

the IOS image is loaded in memory at startup.

Do you check the file on flash with verify comand and you find that is corrupted?

if so you should think of finding a way to replace the internal flash

Hope to help

Giuseppe

mark shuford Mon, 05/17/2010 - 10:34

Thanks for the response Giuseppe. I didn't verify flash but when it wouldn't boot I issued the boot flash: command and it gave an error message stating the the mzip file is corrupt.

Correct Answer
Leo Laohoo Mon, 05/17/2010 - 17:14

Giuseppe's right (+5).  Everytime the switch crashes it creates a file or subdirectory called crashinfo.  If the flash isn't cleaned out and it keeps crashing the switch will go to a state called crash-n-boot loop.  The switch will crash, creates a crashinfo file/subdirectory.  But since there's no more room at the "inn" or flash the switch will crash until someone will have to intervene.  Initially it will look like the IOS is corrupted but it isn't.

To delete subdirectories, use the command "delete /f /r flash:"

Now even if you've cleaned out the flash and you still get this error message about the "mzip files" then as Giuseppe has said, IOS upgrade is the way to go.  You have two avenues of doing this:  you can upgrade the IOS and boot the switch OR you can pump the IOS (if you have enough flash space) and wait for the next crash.

Hope this helps and don't forget to rate useful posts.

mark shuford Tue, 05/18/2010 - 08:41

I had the crashinfo directory with nothing in it. Anyway it crashed again a few hours after reloading the IOS. This time I was logged into the switch when it happened and here is some of the info that was displayed:

Unexpected exception to CPUvector 2000, PC = C08064
-Traceback= C08064 C086CC C13428 C13E38 C13ED4 C184EC C1ED78 C1F194 8F5D6C 8F6178 91A01C 9                                                                   349A8 75E1A8 7583EC


=== Flushing messages (21:28:56 EST Sun Feb 28 1993) ===

Buffered messages: (last 4096 bytes only)
R:6015 0027.0d7b.ab80 00000000
*Feb 28 19:40:37.595:     B:6015 0027.0d7b.ab80 80.86 A:0000 M:1400 H:0200 F:0F00
*Feb 28 19:40:37.595:      T:0000 L:0002 D:0015
*Feb 28 19:40:37.595: STP SW: PROC RX: 0100.0ccc.cccd<-001b.5481.e295 type/len 0032
*Feb 28 19:40:37.595:     encap SNAP linktype sstp vlan 21 len 64 on v21 Gi1/0/12
*Feb 28 19:40:37.595:     AA AA 03 00000C 010B SSTP
*Feb 28 19:40:37.595:     CFG P:0000 V:00 T:00 F:00 R:6015 0027.0d7b.ab80 00000000
*Feb 28 19:40:37.595:     B:6015 0027.0d7b.ab80 80.86 A:0000 M:1400 H:0200 F:0F00
*Feb 28 19:40:37.595:      T:0000 L:0002 D:0015
*Feb 28 19:40:37.595: STP: VLAN0021 rx BPDU: config protocol = rstp, packet from GigabitEt                                                                   hernet1/0/12  , linktype SSTP , enctype 3, encsize 22
*Feb 28 19:40:37.595: STP: enc 01 00 0C CC CC CD 00 1B 54 81 E2 95 00 32 AA AA 03 00 00 0C                                                                    01 0B
*Feb 28 19:40:37.595: STP: Data     0000000000601500270D7BAB8000000000601500270D7BAB808086                                                                   0000140002000F00
*Feb 28 19:40:37.595: STP: VLAN0021 Gi1/0/12:0000 00 00 00 601500270D7BAB80 00000000 60150                                                                   0270D7BAB80 8086 0000 1400 0200 0F00
*Feb 28 19:40:37.595: RSTP(21): Gi1/0/12 repeated msg

The rest of the output has been omitted.

When I rebooted and switch was initializing flash I got the following error:

Base ethernet MAC Address: 00:14:6a:cb:c1:80
Xmodem file system is available.
The password-recovery mechanism is enabled.
Initializing Flash...
flashfs[0]: sector 132, block 111 checksum (0xacd != 0xe2f0)
flashfs[0]: sector 226, block 117 checksum (0x438b != 0xe2f0)
flashfs[0]: sector 234, block 117 checksum (0xb2ae != 0xe2f0)
flashfs[0]: sector 167, block 119 checksum (0x8274 != 0xe2f0)
flashfs[0]: sector 168, block 119 checksum (0xcbfd != 0xe2f0)
flashfs[0]: sector 220, block 119 checksum (0x5c0c != 0xe2f0)
flashfs[0]: sector 223, block 119 checksum (0x8787 != 0xe2f0)
flashfs[0]: sector 229, block 119 checksum (0x76f != 0xe2f0)
flashfs[0]: sector 230, block 119 checksum (0x3be2 != 0xe2f0)
flashfs[0]: sector 115, block 123 checksum (0xd30e != 0xe2f0)
flashfs[0]: sector 12, block 124 checksum (0xb9b7 != 0xe2f0)
flashfs[0]: sector 141, block 124 checksum (0x6706 != 0xe2f0)
flashfs[0]: sector 149, block 124 checksum (0xe412 != 0xe2f0)
flashfs[0]: sector 160, block 124 checksum (0x1f17 != 0xe2f0)
flashfs[0]: sector 161, block 124 checksum (0x3f66 != 0xe2f0)
flashfs[0]: sector 162, block 124 checksum (0x288a != 0xe2f0)
flashfs[0]: sector 171, block 124 checksum (0xdae9 != 0xe2f0)
flashfs[0]: sector 172, block 124 checksum (0xbf92 != 0xe2f0)
flashfs[0]: sector 185, block 124 checksum (0xf481 != 0xe2f0)
flashfs[0]: sector 194, block 124 checksum (0x2c54 != 0xe2f0)
flashfs[0]: sector 196, block 124 checksum (0x1a61 != 0xe2f0)
flashfs[0]: sector 198, block 124 checksum (0x58c3 != 0xe2f0)
flashfs[0]: sector 209, block 124 checksum (0xb29e != 0xe2f0)
flashfs[0]: sector 212, block 124 checksum (0x4e23 != 0xe2f0)
flashfs[0]: sector 214, block 124 checksum (0x8fb6 != 0xe2f0)
flashfs[0]: sector 217, block 124 checksum (0x2391 != 0xe2f0)
flashfs[0]: sector 218, block 124 checksum (0xd61e != 0xe2f0)
flashfs[0]: sector 219, block 124 checksum (0x8e41 != 0xe2f0)
flashfs[0]: sector 220, block 124 checksum (0xaac1 != 0xe2f0)
flashfs[0]: sector 226, block 124 checksum (0x2819 != 0xe2f0)
flashfs[0]: sector 227, block 124 checksum (0x9425 != 0xe2f0)
flashfs[0]: sector 228, block 124 checksum (0xd704 != 0xe2f0)
flashfs[0]: sector 230, block 124 checksum (0x5d67 != 0xe2f0)
flashfs[0]: sector 232, block 124 checksum (0xd621 != 0xe2f0)
flashfs[0]: sector 233, block 124 checksum (0xa176 != 0xe2f0)
flashfs[0]: sector 234, block 124 checksum (0x7926 != 0xe2f0)
flashfs[0]: sector 235, block 124 checksum (0x7ab4 != 0xe2f0)
flashfs[0]: inconsistent sector list, fileid 6, parent_fileid 0
Warning: Trying to free already-freed memory!!!
flashfs[0]: Trying to read a sector not in the flash device!.
flashfs[0]: block_num (2047) not in range between 0 and 124.
...done Initializing Flash.
Boot Sector Filesystem (bs) installed, fsid: 3

The system has been interrupted, or encountered an error
during initializion of the flash filesystem.  The following
commands will initialize the flash filesystem, and finish
loading the operating system software:

    flash_init

Correct Answer
Leo Laohoo Wed, 05/19/2010 - 00:14

You've unfortunately hit a hole-in-one in terms of IOS bug and it's a very bad one too.  IOS upgrade is the only way to fix it.

Message was edited by: leolaohoo

mark shuford Wed, 05/19/2010 - 04:56

I did upgrade the IOS after that crash. Thanks leolaohoo for the help. Now I need to do the upgrade at 25 other sites because they are using the same IOS.

Actions

This Discussion