communication failure

  • Using OMV 3.0.86 (Erasmus).


    I get "communication failure" error multiple times in Storage/File systems, after starting filesystem creation on 500 Gb HD, ext4 fs.
    I get this error on any page at the moment.


    Why does it happen?
    How can I debug it (I'm PHP developer)?

    • Offizieller Beitrag

    This is a result of the filesystem command that engined (written in php) not completing in time that extjs (javascript) allows. Is the drive failing or being heavily used? Have a look at this code

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Currently I'm formatting /dev/sdc.

    Code
    root@omvnas:~# ps -aux | grep mkfs.ext4
    root      1528  0.0  0.0  12728  2160 pts/0    S+   18:03   0:00 grep mkfs.ext4
    root     17905  0.5  0.0  23800  6652 ?        D    17:05   0:20 mkfs.ext4 -b 4096 -m 0 -E lazy_itable_init=0,lazy_journal_init=0 -O 64bit -L sr2 /dev/sdc1

    There is no other reason for communication error.

  • Maybe the following code throws exception, and doesn't return any info

    Code
    throw new \OMV\ExecException($cmd->getCommandLine(), $output,
    					$exitStatus);


    I don't find any "communication failure" string in the file above.


    The only string I found:
    https://github.com/openmediava…nication+failure%22&type=


    Code
    openmediavault (3.0.20) stable; urgency=low
    
    
      * Update locales.
      * Various improvements and bugfixes.
    * Mantis 0001522: Communication failure popup clicking
    • Offizieller Beitrag

    Currently I'm formatting /dev/sdc.


    There is no other reason for communication error.

    High load on the system can cause the problem. What kind of system?

    I don't find any "communication failure" string in the file above.

    The failure comes from the javascript side. I just pointed to the php code that isn't returning info because the command hasn't completed.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • High load on the system can cause the problem. What kind of system?

    The failure comes from the javascript side. I just pointed to the php code that isn't returning info because the command hasn't completed.

    System is an MSI server motherboard with 16 Gb memory, dual AMD OPTERON processor.
    It is spare now, so nothing memory or processor intensive tasks are running on it.


    Should not have performance problems.
    The problem clearly comes from OMV 3.


    The problem is that javascript doesn't give any information, what script was executed.
    Is debug mode implemented in OMV at all?


    Sent with Tapatalk from my SM-N9005

  • I have the same problem. Please see my posts in this thread.


    I get it when a ZFS scrub is running in the background, which generates a heavy disk load. Or when the drives are in standby and need to be started first. There are several other threads where people report this issue, when they are doing disk related tasks (File system actions, S.M.A.R.T. a.s.o)


    I have filed a bug report at bugtracker some days ago, but no reply so far (Absolutely no criticism!). Therefore I think there is no simple solution for this behavior.

    OMV 3.0.100 (Gray style)

    ASRock Rack C2550D4I C0-stepping - 16GB ECC - 6x WD RED 3TB (ZFS 2x3 Striped RaidZ1) - Fractal Design Node 304 -

    3x WD80EMAZ Snapraid / MergerFS-pool via eSATA - 4-Bay ICYCube MB561U3S-4S with fan-mod

    Einmal editiert, zuletzt von cabrio_leo ()

    • Offizieller Beitrag

    Should not have performance problems.
    The problem clearly comes from OMV 3.

    Clearly? What is the problem then? The error is a javascript error which means the client is involved as well. This means the browser, networking, and the server are all involved. Why doesn't everyone get the error? Why is the system performing a relatively high resource action in pretty much all of the situations I have seen this involved with? I get that multiple people are seeing this but it is not easy to diagnose especially when it can't be recreated on purpose.


    Is debug mode implemented in OMV at all?

    Yep and it works very well.
    monit stop omv-engined
    omv-engined -df

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Clearly? What is the problem then? The error is a javascript error which means the client is involved as well. This means the browser, networking, and the server are all involved. Why doesn't everyone get the error? Why is the system performing a relatively high resource action in pretty much all of the situations I have seen this involved with? I get that multiple people are seeing this but it is not easy to diagnose especially when it can't be recreated on purpose.


    Yep and it works very well.
    monit stop omv-engined
    omv-engined -df

    At least 2 of 5 HDDs in the server has bad sectors (content moved to reserved space). At least one doesn't pass short SMART test.
    This might be the reason of the error, and the reason why doesn't happen for everybody.



    Can I increase somewhere the javascript gui response timeout?


    I will try the debug mode later.


    Sent with Tapatalk from my SM-N9005

    • Offizieller Beitrag

    Can I increase somewhere the javascript gui response timeout?

    Probably but I don't know where it is. I think that is avoiding the problem as well. Failing drives mean commands aren't returning in time and that is a timeout not set by OMV.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Probably but I don't know where it is. I think that is avoiding the problem as well. Failing drives mean commands aren't returning in time and that is a timeout not set by OMV.

    I think, that a failing drive should not cause failure of the GUI.


    Furthermore, the GUI should warn & display the drive failure.


    Sent with Tapatalk from my SM-N9005

    • Offizieller Beitrag

    I think, that a failing drive should not cause failure of the GUI.


    Furthermore, the GUI should warn & display the drive failure.

    Great thought but OMV doesn't do magic. Any command run from the command line to query a drive's status will be very slow if it has failed in certain ways. OMV uses these same commands. A commercial NAS vendor can show warnings/failures like this because they always have the same hardware. OMV does not have that luxury. Sleeping drives can also appear to be failed if not given enough time to start.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Great thought but OMV doesn't do magic. Any command run from the command line to query a drive's status will be very slow if it has failed in certain ways. OMV uses these same commands. A commercial NAS vendor can show warnings/failures like this because they always have the same hardware. OMV does not have that luxury. Sleeping drives can also appear to be failed if not given enough time to start.

    I think it is development decision how to react, when timeout happens.
    OMV developer decided to give a fatal error, with a dialog window to be confirmed.


    I think there could be another way to handle such situation.


    Android gives a warning window, when an app is unresponsive, giving some options:
    - wait
    - report
    - close app


    OMV could handle similar way.
    If GUI doesn't get response from the server, should display a dialog window with multiple choices:
    - wait (adds +1 minute timeout before next warning dialog shows)
    - report (do so automatic debug information collection, and report to the developers). The report way should also make possible to preview the debug informations to be sent.
    - close dialog.


    I think this would be a better way to handle such timeout situation.




    Sent with Tapatalk from my SM-N9005

  • Further possibility:
    If I execute a task from GUI, it should:
    - pass command from GUI to server
    - server starts execution, and writes result buffer continuously to a session log file until it is finished successfully or finished with error.
    - GUI should read session log file, when requested by the user from GUI (by pressing "Command log" button on Storage/File systems page) or display the finished command output in an informational dialog on any current page.


    This is how it should work, IMHO.


    Sent from my GT-I9305 using Tapatalk.

    • Offizieller Beitrag

    If I execute a task from GUI, it should:
    - pass command from GUI to server
    - server starts execution, and writes result buffer continuously to a session log file until it is finished successfully or finished with error.
    - GUI should read session log file, when requested by the user from GUI (by pressing "Command log" button on Storage/File systems page) or display the finished command output in an informational dialog on any current page.

    This is what it currently does. BUT if the Linux command (not written by OMV) does not return any output (which is common), the OMV web interface shows the communication error (not fatal in my opinion since it will recover).

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • I have the same problem: 36 disks in the system and the normal smart disk listing does take too long and the javascript throws an error after about 20 seconds. The RPC would have been finished after 32 seconds.
    (@ryecoaaron Thank you for the help getting debug info)


    I tried to fiend the rpc js timeout settings but without luck. Anyone who knows where i can find them ?


    Cheers

    • Offizieller Beitrag

    The only timeout I can find is the HTTP timeout request in /usr/share/php/openmediavault/globals.inc and it is set to 60 seconds.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Just wanted to say that I got the same problem. I get the communication error or time-out error very often and can't enable/disable services anymore because of it which is very frustrating.


    Is there maybe a way to enable/disable plugins/services via putty?

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!