• .?lo files disappearing error 11

    From deon@1:103/705 to Digital Man on Mon Jun 2 17:57:53 2025
    Howdy,

    I've just discovered something weird, that has been happening for a month or two. Probably since I moved from x86_64 to aarch64.

    When outbound FTN mail is created, sbbsecho is correctly placing the .?lo file in the zone oubound dir, and triggering binkit to send it.

    It seems when binkit calls the remote, when it goes to open the .?lo file to see what to send, its gone. I can re-produce this pretty easily, its happening for all my FTN networks.

    As a consequence, no mail gets sent.

    Here is an except from syslog:

    Jun 02 15:42:47 ? local1.info synchronet: evnt FIDOOUT Running native timed event: /opt/sbbs/exec/sbbsecho -ni
    Jun 02 15:42:49 ? local1.info synchronet: evnt FIDOOUT Timed event: '/opt/sbbs/exec/sbbsecho -ni' returned 0
    Jun 02 15:42:50 ? local1.info synchronet: evnt BBS Events Semaphore signaled for Timed Event: BINKOUT
    Jun 02 15:42:50 ? local1.info synchronet: evnt BINKOUT Running native timed event: ?binkit
    Jun 02 15:42:50 ? local1.info synchronet: evnt BINKOUT BinkIT/2.41 invoked with options:
    Jun 02 15:42:51 ? local1.info synchronet: evnt BINKOUT Attempting callout for 12:1/1@dovenet, file: /opt/sbbs/fido/outbound.00c/00010001.dlo
    Jun 02 15:42:51 ? local1.info synchronet: evnt BINKOUT JSBinkP/4 callout to 12:1/1@dovenet started
    Jun 02 15:42:51 ? local1.info synchronet: evnt BINKOUT Connecting to 12:1/1@dovenet at clrghouz.bbs.dege.au:24554
    Jun 02 15:43:00 ? local1.info synchronet: evnt BINKOUT Will encrypt session. Jun 02 15:43:00 ? local1.info synchronet: evnt BINKOUT Authentication successful: secure
    Jun 02 15:43:28 ? local1.err synchronet: evnt BINKOUT Error 11 opening FLO file: /opt/sbbs/fido/outbound.00c/00010001.dlo
    Jun 02 15:43:31 ? local1.info synchronet: evnt BINKOUT Timed event: '?binkit' returned 0

    I just rebuilt sbbs today, as my last build was about a month or two old - still the same.

    Why the error 11, and what would cause the file to disappear?

    If I create the .?lo manually, and send it interactively with jsexec binkit -l ... , the file doesnt disappear (until the mail is sent), and the mail gets transfered correctly.

    I think this is related to me switching to aarch64. I'm using the same config, and the same build process as I have always used.


    ...ëîåï

    ---
    þ Synchronet þ AnsiTEX bringing back videotex but with ANSI
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Digital Man@1:103/705 to deon on Mon Jun 2 14:45:38 2025
    Re: .?lo files disappearing error 11
    By: deon to Digital Man on Mon Jun 02 2025 05:57 pm

    Howdy,

    I've just discovered something weird, that has been happening for a month or two. Probably since I moved from x86_64 to aarch64.

    When outbound FTN mail is created, sbbsecho is correctly placing the .?lo file in the zone oubound dir, and triggering binkit to send it.

    It seems when binkit calls the remote, when it goes to open the .?lo file to see what to send, its gone. I can re-produce this pretty easily, its happening for all my FTN networks.

    As a consequence, no mail gets sent.

    Here is an except from syslog:

    Jun 02 15:42:47 ? local1.info synchronet: evnt FIDOOUT Running native timed event: /opt/sbbs/exec/sbbsecho -ni
    Jun 02 15:42:49 ? local1.info synchronet: evnt FIDOOUT Timed event: '/opt/sbbs/exec/sbbsecho -ni' returned 0
    Jun 02 15:42:50 ? local1.info synchronet: evnt BBS Events Semaphore signaled for Timed Event: BINKOUT
    Jun 02 15:42:50 ? local1.info synchronet: evnt BINKOUT Running native timed event: ?binkit
    Jun 02 15:42:50 ? local1.info synchronet: evnt BINKOUT BinkIT/2.41 invoked with options:
    Jun 02 15:42:51 ? local1.info synchronet: evnt BINKOUT Attempting callout for 12:1/1@dovenet, file: /opt/sbbs/fido/outbound.00c/00010001.dlo
    Jun 02 15:42:51 ? local1.info synchronet: evnt BINKOUT JSBinkP/4 callout to 12:1/1@dovenet started
    Jun 02 15:42:51 ? local1.info synchronet: evnt BINKOUT Connecting to 12:1/1@dovenet at clrghouz.bbs.dege.au:24554
    Jun 02 15:43:00 ? local1.info synchronet: evnt BINKOUT Will encrypt session. Jun 02 15:43:00 ? local1.info synchronet: evnt BINKOUT Authentication successful: secure
    Jun 02 15:43:28 ? local1.err synchronet: evnt BINKOUT Error 11 opening FLO file: /opt/sbbs/fido/outbound.00c/00010001.dlo
    Jun 02 15:43:31 ? local1.info synchronet: evnt BINKOUT Timed event: '?binkit' returned 0

    I just rebuilt sbbs today, as my last build was about a month or two old - still the same.

    Why the error 11, and what would cause the file to disappear?

    /usr/include/asm-generic/errno-base.h:#define EAGAIN 11 /* Try again */

    I'm guessing that means the file was opened for exclusive access or the file system was doing something with the file. It's not saying the file "is gone".

    If I create the .?lo manually, and send it interactively with jsexec binkit -l ... , the file doesnt disappear (until the mail is sent), and the mail gets transfered correctly.

    I think this is related to me switching to aarch64. I'm using the same config, and the same build process as I have always used.

    It sounds like BinkIt should peform an open-retry when getting that specific open error. Submit an issue at gitlab.synchro.net for someone to follow-up on this?
    --
    digital man (rob)

    Sling Blade quote #2:
    Karl (re: killing Doyle): I hit him two good whacks in the head with it.
    Norco, CA WX: 79.3øF, 53.0% humidity, 12 mph WSW wind, 0.00 inches rain/24hrs --- SBBSecho 3.27-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From deon@1:103/705 to Digital Man on Tue Jun 3 08:15:46 2025
    Re: .?lo files disappearing error 11
    By: Digital Man to deon on Mon Jun 02 2025 02:45 pm

    Howdy,

    Why the error 11, and what would cause the file to disappear?

    /usr/include/asm-generic/errno-base.h:#define EAGAIN 11 /* Try again */

    I'm guessing that means the file was opened for exclusive access or the file system was doing something with the file. It's not saying the file "is gone".

    Oh, I'm telling you the file is gone.

    My outbound directory is left with a bunch of packets and no .?lo file.

    If I "watch ls -al outbound.xxx/", I can see sbbsecho create the packet, create the .?lo file.

    As soon as binkit goes to send it, it disappears, and the Error 11 is reported at the same time.

    I have mail going out frequently (at least every hour), and it is happening every time.

    "Try again" doesnt make sense for this - as the file is not there..

    It sounds like BinkIt should peform an open-retry when getting that specific open error. Submit an issue at gitlab.synchro.net for someone to follow-up on this?

    OK.



    ...ëîåï

    ---
    þ Synchronet þ AnsiTEX bringing back videotex but with ANSI
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From Digital Man@1:103/705 to deon on Mon Jun 2 15:34:14 2025
    Re: .?lo files disappearing error 11
    By: deon to Digital Man on Tue Jun 03 2025 08:15 am

    Re: .?lo files disappearing error 11
    By: Digital Man to deon on Mon Jun 02 2025 02:45 pm

    Howdy,

    Why the error 11, and what would cause the file to disappear?

    /usr/include/asm-generic/errno-base.h:#define EAGAIN 11
    /* Try again */

    I'm guessing that means the file was opened for exclusive access or the file system was doing something with the file. It's not saying the file "is gone".

    Oh, I'm telling you the file is gone.

    My outbound directory is left with a bunch of packets and no .?lo file.

    If I "watch ls -al outbound.xxx/", I can see sbbsecho create the packet, create the .?lo file.

    As soon as binkit goes to send it, it disappears, and the Error 11 is reported at the same time.

    Most likely, BinkIt is deleting the file itself immediately after that error is logged.

    I would expect to see a "Deleted file: /path/to/flo-file" log message in that case, however, and you didn't show that, so that's suspicious.

    I have mail going out frequently (at least every hour), and it is happening every time.

    "Try again" doesnt make sense for this - as the file is not there..

    I suspect the file *is* there at the moment that BinkIt is trying to open it and is logging that error message.

    Looking at the script source (which mind you, I did not write), the File.open() method is be used to open FLO files and in shareable mode, so the file should be automatically retry the open and only return that error after several failed open attempts (usually a noticable delay of several seconds). The 28 second gap between log messages the log output you shared suggests that these open retries are indeed happening.

    Is it possible SBBSecho or another process still has this FLO file open at that moment?

    It sounds like BinkIt should peform an open-retry when getting that specific open error. Submit an issue at gitlab.synchro.net for someone to follow-up on this?

    OK.

    Thanks,
    --
    digital man (rob)

    Synchronet/BBS Terminology Definition #56:
    NNTP = Network News Transfer Protocol
    Norco, CA WX: 79.4øF, 49.0% humidity, 14 mph WNW wind, 0.00 inches rain/24hrs --- SBBSecho 3.27-Linux
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
  • From deon@1:103/705 to Digital Man on Tue Jun 3 10:47:30 2025
    Re: .?lo files disappearing error 11
    By: Digital Man to deon on Mon Jun 02 2025 03:34 pm

    Most likely, BinkIt is deleting the file itself immediately after that error is logged.

    I would expect to see a "Deleted file: /path/to/flo-file" log message in that case, however, and you didn't show that, so that's suspicious.

    Yes, no mention of "Deleted file:" is shown in syslog (from the event server). I do see that when I run it interactively though.

    Is it possible SBBSecho or another process still has this FLO file open at that moment?

    I dont believe so.

    My test FTN area (zone 10 "private") is only used by me - and if I am the only person posting a message, sbbsecho does it thing.

    Here is an except from that time in question:

    2025-06-02 15:42:49 Finalizing outbound packet from 12:1/2 to 12:1/1: /opt/sbbs/temp/sbbsecho/683d39d8.pkt
    2025-06-02 15:42:49 Node (12:1/1) successfully locked via: ../fido/outbound.00c/00010001.bsy
    2025-06-02 15:42:49 Sending packet (/opt/sbbs/temp/sbbsecho/683d39d8.pkt, 4.5KB) from 12:1/2 to 12:1/1
    2025-06-02 15:42:49 Moving packet for 12:1/1: /opt/sbbs/temp/sbbsecho/683d39d8.pkt to ../fido/outbound.00c/683d39d8.pkt
    2025-06-02 15:42:49 Deleting /opt/sbbs/temp/sbbsecho/683d39d8.pkt (from line 3037)
    2025-06-02 15:42:49 File (../fido/outbound.00c/683d39d8.pkt, 4.5KB) for 12:1/1 added to BSO/FLO file: ../fido/outbound.00c/00010001.dlo
    2025-06-02 15:42:49 Touching outgoing semfile: ../data/binkout.now
    2025-06-02 15:42:49 Writing 17 areas to ../data/badareas.lst
    2025-06-02 15:42:49 Deleting ../fido/outbound.00c/00010001.bsy (from line 3200) 2025-06-02 15:42:49 Deleting /opt/sbbs/ctrl/sbbsecho.bsy (from line 3206) 2025-06-02 15:42:49 SBBSecho (PID 3545) exiting with error level 0, Packets(0 imported, 1 sent), EchoMail(0 imported, 2 exported)

    2025-06-02 15:43:32 SBBSecho 3.27-Linux master/7276d373a Jun 01 2025 GCC 10.2.1 (PID 3565) invoked with options: -ce
    2025-06-02 15:43:32 Configured: 7 archivers, 57 linked-nodes, 13 echolists

    From this, I dont believe sbbsecho was running (and working on that .?lo) file at the same time binkit was reading it. (Binkit started at 15:42:50, when sbbsecho finished at 15:42:49.)


    ...ëîåï

    ---
    þ Synchronet þ AnsiTEX bringing back videotex but with ANSI
    * Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)