Forum Discussion

Thomas_McLean_1's avatar
Thomas_McLean_1
Icon for Nimbostratus rankNimbostratus
Sep 09, 2011

F5 BIG-IP configuration backup

All,

 

 

We're having an issue with our BIG-IP LTM 3900 in which we cannot backup our configuration file as we are going to be doing an upgrade to our environment...we are receiving this error:

 

 

Does anyone know how to rectify this error?

 

 

Thanks in advance,

 

 

Thomas.

 

 

 

Operation Status Saving active configuration... mkdir: cannot create directory `/var/tmp/cert_temp': Read-only file system /tmp/cs_save_pre_script: line 41: cd: /var/tmp/cert_temp: No such file or directory /tmp/cs_save_pre_script: line 61: cd: /var/tmp/cert_temp: No such file or directory /tmp/cs_save_pre_script: line 77: cd: /var/tmp/cert_temp: No such file or directory mkdir: cannot create directory `/var/tmp/gtm_tmp': Read-only file system mkdir: cannot create directory `/var/tmp/gtm_tmp': Read-only file system /tmp/cs_save_pre_script: line 105: cd: /var/tmp/gtm_tmp/gtm: No such file or directory /tmp/cs_save_pre_script: line 108: cd: /var/tmp/gtm_tmp/gtm: No such file or directory mkdir: cannot create directory `/var/tmp/gtm_tmp': Read-only file system /tmp/cs_save_pre_script: line 116: cd: /var/tmp/gtm_tmp/big3d: No such file or directory mkdir: cannot create directory `/var/tmp/em_db_temp': Read-only file system /var/tmp/cert_temp: Directory does not exist Error creating temporary directory /var/tmp/im7606 WARNING:There are error(s) during saving. Not everything was saved. Be very careful when using this saved file! Error creating package Error during config save. BIGpipe parsing error: 01020001:3: The requested operation failed.
management
monitoring

7 Replies

Sort By
  • nitass's avatar
    nitass
    Icon for Employee rankEmployee
    Sep 09, 2011
    what HDD firmware version are you using?

     

    there is the firmware upgrade in EUD.

     

     

    also, rebooting unit would fix it.
  • Thomas_McLean_1's avatar
    Thomas_McLean_1
    Icon for Nimbostratus rankNimbostratus
    Sep 09, 2011
    Hi nitass,

     

     

    BIG-IP10.2.01755.1 <-- i'm not sure exactly where I would be able to find the HDD fw.

     

     

    it's a BIG-IP 3600 system. Pretty new to it as we put it in and basically forgot about it now there's need to move it from our old environment to our new environment and I don't want to move it without having a copy of the config.

     

     

    Thanks again
  • nitass's avatar
    nitass
    Icon for Employee rankEmployee
    Sep 09, 2011
    as far as i know, upgrading HDD firmware and rebooting unit would fix this read-only file system issue.

     

     

    sol12780: Hard drive shipped with BIG-IP 3900 and EM4000 platforms before September 2010 may exhibit failure symptoms due to firmware issue

     

    http://support.f5.com/kb/en-us/solutions/public/12000/700/sol12780.html
  • Thomas_McLean_1's avatar
    Thomas_McLean_1
    Icon for Nimbostratus rankNimbostratus
    Sep 09, 2011
    nitass, again thanks for your quick reply - I am one of the affected so will need to schedule in downtime and carry out diags and an upgrade of our firmware.

     

     

    Let's hope it isn't too painful really :)

     

     

    Cheers,

     

     

    Thomas.
  • WeaverJK's avatar
    WeaverJK
    Icon for Nimbostratus rankNimbostratus
    Apr 19, 2017

    Potentially, a directory structure is full.

     

    Experienced the same symptoms when attempting to create an archive (backup) file. In our situation...

     

    Issue: df showed that /shared was 100% used.

     

    Cause: mcpd.out file was 55 GB

     

    Root Cause: DISA STIGs suggest setting the mcpd logging level to debug.

     

    Side Note: Many IA folks believe that the security technical implementation guides are not guidelines but hard requirements. Representatives from both DISA and FSS stated that STIGs are suggestions and are not requirements.

     

    Temporary Solution: transferred Traffic Groups to alternate F5, then rebooted the F5 experiencing the problem. Note that I've previously tried other methods of resetting the mdpd.out file, only to have each require a reboot.

     

    Long-term solution: Option A) create a cron job to perform log rotation function; Option B) revert to default logging level of mcpd

     

    Please share if you've found differing (and preferably better) permanent solution.

     

  • Hamish's avatar
    Hamish
    Icon for Cirrocumulus rankCirrocumulus
    Apr 19, 2017

    A blast from the past... 6 years ago.. However you can usually empty an open file by cat'ing dev/null to it. e.g.

     

    cat /dev/null > FILENAME

     

    No reboot necessary.

     

    H

     

  • WeaverJK's avatar
    WeaverJK
    Icon for Nimbostratus rankNimbostratus
    Apr 19, 2017

    Understood, Hamish.

     

    When this issue presented, I reviewed a list of about three options, including cat'ing to dev/null.

     

    Here are my notes from that issue: Internet posts identify concerns with having to reassign file permissions.

     

    > /shared/tmp/mcpd.out did not produce the desired resutls. The file is still 41-GB in size. The “df -h” shows /shared as only 9% used now (good). If, however, the mcpd.out file is viewed (such as with more), the first 41-GB of the file is blank. In order to see the current entries, tail must be used.

     

    What appeared to work (but didn't) was this: truncate /shared/tmp/mcpd.out --size 0

     

    My notes end here. What I recall, however, is that the linux inodes were broken and the recover was to reboot.

     

    I did not test cat. Do you foresee this breaking the inodes or causing any other issues, such as with permissions?