Profile Details

Toggle Sidebar
Loading cover... Drag cover to reposition
Recent updates

  • So for now as far as I understand I should set unused interfaces as internal?

    Yes, that's better than how they are configured now... You are susceptible with all these ping failures to syswatch restarting and dropping connections temporarily...
    Reading your first append again I see you are really on 7.3 now... Can you run "ip route show table 250" when you have updated your network.
    Table 250 may not have shown in the output you provided as a result of the two unused interfaces... or ???
    Another thought - what does "netstat -r" show? (after re-config)

  • Nick is spot-on - never have configured unused WAN ports with Multi-WAN. Also agree re. LAN ports on same sub-net - in this case the lower of the two IPs is the one that the system will attempt to use...

    The output from "ip route show table all | awk '/default/ && /table/' | egrep -v 'prohibit|error|proto'" has me perplexed - there is no table 250...
    Seems to indicate yours is missing or you are really running ClearOS 6.x. (you indicated 7.2) What is the output of "ip route show table 250" ?

    Example output from ClearOS 7.x multi-wan system

    Example from ClearOS 6.x multi-wan system

  • Tony Ellis
    Tony Ellis's reply was accepted as an answer

    Re: Install clearOS on a USB flash drive?

    I'll not go into detail - but to be successful you will need to use several techniques to minimize writes to the USB flash
    These include things such as (no particular order - just what comes to mind)

    1) reducing the amount of logging - edit /etc/rsyslog.conf - for logging config options for other programs
    2) creating a memory drive for the remaining logs (lost at power-down/reboot) - optional
    3) if using ext filesystems - specify noatime option in /etc/fstab (not necessary for xfs)
    4) don't use programs that are memory write intensive such as the squid proxy disable others eg updatedb
    5) set /proc/sys/vm/swappiness to zero
    6) install plenty of memory to reduce swapping to practically zero
    7) use a really good quality USB flash drive - no cheapies

    I'm sure there are plenty more hints on the web if you do a search

    As for is it possible? - can your BIOS boot from a USB drive?
    Safe/Wise? That's your risk assessment. As a minimum once you have completed the install use "dd" to make a complete copy of your USB flash drive onto another identical USB drive. Do NOT copy the drive running the current system - for instance use the install USB drive and use the option to get a command line.. This requires 3 USB drives...
    1) The one you booted from
    2) the one you are copying - completed installation (using dd) FROM
    3) the one you are copying TO
    Make sure you know which is which :-)

    Should give you food for thought...

  • Yuk! - both interfaces in trouble - what percentage of the time are they in that state?

    As a test are you able to set both modem/routers on your external interfaces to NAT mode and then set them up together with your ClearOS box to provide private address range static ips to the ClearOS multiwan interfaces. Each ClearOS external interface will need to be on a different sub-net and also each different to all of your internal sub-nets. Be sure to check the default routes and DNS servers are correctly setup.

    If you are able to do this can you provide the output from the commands below... if you cannot do the test - can you show the results anyway...

    # ip -4 address
    # ip route show table all | awk '/default/ && /table/' | egrep -v 'prohibit|error|proto'

    and an extract from syswatch from the time you run "systemctl restart syswatch.service" until you either we have "info: system - heartbeat..." or a few lines of ping failures.

  • Al - can you append the entries for the kernel update from "/var/log/messages" when your kernel update actually took place - see the discussion with MCB as to what we are looking for...

  • Just remembered this... Provided you can boot from USB - here's something much more durable...

    https://arstechnica.com/gadgets/2016/03/western-digital-makes-a-46-314gb-hard-drive-just-for-the-raspberry-pi/

  • I'll not go into detail - but to be successful you will need to use several techniques to minimize writes to the USB flash
    These include things such as (no particular order - just what comes to mind)

    1) reducing the amount of logging - edit /etc/rsyslog.conf - for logging config options for other programs
    2) creating a memory drive for the remaining logs (lost at power-down/reboot) - optional
    3) if using ext filesystems - specify noatime option in /etc/fstab (not necessary for xfs)
    4) don't use programs that are memory write intensive such as the squid proxy disable others eg updatedb
    5) set /proc/sys/vm/swappiness to zero
    6) install plenty of memory to reduce swapping to practically zero
    7) use a really good quality USB flash drive - no cheapies

    I'm sure there are plenty more hints on the web if you do a search

    As for is it possible? - can your BIOS boot from a USB drive?
    Safe/Wise? That's your risk assessment. As a minimum once you have completed the install use "dd" to make a complete copy of your USB flash drive onto another identical USB drive. Do NOT copy the drive running the current system - for instance use the install USB drive and use the option to get a command line.. This requires 3 USB drives...
    1) The one you booted from
    2) the one you are copying - completed installation (using dd) FROM
    3) the one you are copying TO
    Make sure you know which is which :-)

    Should give you food for thought...

  • Yes - that's nothing more than just the yum.log - what we want is the content from /var/log/messages for the kernel install at about that time...
    here's what was found here to illustrate. Sorry if wasn't clear - "check the others for anything during that same time-frame" meant check the other logs in /var/log for entries about that same time.

    As you can see - the last item was the initramfs created by dracut

    Now what is in yours???

  • Al


    my log partition of 8gb was completely full

    Separate partition just for logs? (/var/log/messages), for var? or in fact was this really the root partition? Please clarify - "df -Th"

  • Hmm... would have just rebooted into an older kernel first and simply used yum to reinstall the new kernel...

    Currently running kernel 3.10.0-514.2.2.v7.x86_64. Just now used yum to install the newer kernel - but not yet rebooted - have tests running don't want to disrupt..
    However looks good- all five files are there

    Small section of /etc/grub2.cfg

    So like Nick should have no problems...
    Here's an idea of the total space used - now 3 kernels - was 2

    It would appear that your install aborted for some reason - did you check if there was there any indication in the logs at the time of install of a failure? /var/log/yum.log will show when it was installed so you can check the others for anything during that same time-frame...