Hello There,

I am contining to have a problem with stale or undeleted busy files. I have decided to use the unprefered option of setting up an event to periodically
use FIDOPOLL to kill all busy files. As it looks like FIDOPOLL doesn't always delete the busy file. Looking at the logs seems to indicate the the FIDOPOLL connection doesn't always end completely or correctly while at other times in the log the connection does complete with the same remote system. Curreently "My" Mystic only connects to two remote systems. A BBBS system and a BinkD system. My presumption is that either BinkD isn't completing the connection ORFIDOPOLL isn't detecting the connection being completed and then doesn't remove the busy file.

This last busy period lasted the better part of 24 hours. I have since added
a Mystic event to delete all busy files once every 12 hours until a better solution to the busy is sue can be found.


Sep 18 00:41:01 Polling BINKP node 21:1/128 (Jeff Smith) by IPV4
Sep 18 00:41:01 Connecting to bbs.ouijabrd.net
Sep 18 00:41:01 Using address 65.103.12.161
Sep 18 00:41:01 Connected IPV4 to 65.103.12.161
Sep 18 00:41:01 S: NUL SYS The Twisted Balloon
Sep 18 00:41:01 S: NUL ZYZ Jeff Smith
[...]
Sep 18 00:41:01 R: ADR 1:282/1031.0@fidonet 1:14/5.0@fidonet 21:1/128.0@fsxnet
1:14/0.0@fidonet 1:299/0.0@fidonet 1:300/0.0@fidonet 1:290/0.0@fidonet Sep 18 00:41:02 S: PWD
Sep 18 00:41:02 R: OK
Sep 18 00:41:02 S: NUL QSIZE 0 files 0 bytes
Sep 18 00:41:02 R: EOB
Sep 18 00:41:02 S: EOB
Sep 18 00:41:02 Session ended (0 sent, 0 rcvd, 0 skip)

Sep 18 00:42:01 FIDOPOLL Version 1.12 A43 2019/03/02
Sep 18 00:42:01 Looking up address "1:3634/12" in nodelist
Sep 18 00:42:01 Scanning 1:3634/12
Sep 18 00:42:01 Queued 0 files (0 bytes) to 1:3634/12
[...]
Sep 18 00:42:01 R: NUL TIME Wed, 18 Sep 2019 01:42:01 -0400
Sep 18 00:42:01 Info TIME Wed, 18 Sep 2019 01:42:01 -0400
Sep 18 00:42:01 R: NUL VER binkd/1.1a-99/Linux binkp/1.1
Sep 18 00:42:01 Mailer binkd/1.1a-99/Linux binkp/1.1
Sep 18 00:42:01 R: ADR 1:3634/12@fidonet 1:3634/0@fidonet 1:123/0@fidonet
1:18/0@fidonet 1:1/120@fidonet 432:1/139@vkradio

----------------- MUTIL v1.12 A43 2019/03/02 Wed, Sep 18 2019 (loglevel 2)
+ Sep 18 00:41:02 Startup using import.ini
+ Sep 18 00:41:02 Process: Importing EchoMail
+ Sep 18 00:41:02 Waiting for BUSY nodes
+ Sep 18 00:41:02 Importing 5d81bf58.pkt (1:3634/12 to 1:14/6)
+ Sep 18 00:41:02 Results: 5 echo, 0 net, 0 dupes, 0 tossed in 0.04s
+ Sep 18 00:41:02 Shutdown Normal (0)

----------------- MUTIL v1.12 A43 2019/03/02 Wed, Sep 18 2019 (loglevel 2)
+ Sep 18 00:42:02 Startup using import.ini
+ Sep 18 00:42:02 Process: Importing EchoMail
+ Sep 18 00:42:02 Waiting for BUSY nodes
+ Sep 18 00:43:02 Results: Cannot import. Some nodes are BUSY in 60.07s
! Sep 18 00:43:02 Status: FATAL
+ Sep 18 00:43:02 Shutdown Normal (0)

----------------- MUTIL v1.12 A43 2019/03/02 Wed, Sep 18 2019 (loglevel 2)
+ Sep 18 01:12:03 Startup using import.ini
+ Sep 18 01:12:03 Process: Importing EchoMail
+ Sep 18 01:12:03 Waiting for BUSY nodes
+ Sep 18 01:13:03 Results: Cannot import. Some nodes are BUSY in 60.07s
! Sep 18 01:13:03 Status: FATAL
+ Sep 18 01:13:03 Shutdown Normal (0)


Jeff

--- Mystic BBS v1.12 A43 2019/03/02 (Linux/64)
* Origin: The Twisted Balloon - bbs.twisted-balloon.org (1:14/6)

On 18 Sep 2019 at 08:17p, Jeff Smith pondered and said...

Hello There,

Hi :)

I am contining to have a problem with stale or undeleted busy files. I

Hmm

have decided to use the unprefered option of setting up an event to periodically use FIDOPOLL to kill all busy files. As it looks like FIDOPOLL doesn't always delete the busy file. Looking at the logs seems
to indicate the the FIDOPOLL connection doesn't always end completely or correctly while at other times in the log the connection does complete with the same remote system. Curreently "My" Mystic only connects to two remote systems. A BBBS system and a BinkD system. My presumption is that either BinkD isn't completing the connection ORFIDOPOLL isn't detecting the connection being completed and then doesn't remove the busy file.

Using fidopoll killbusy all is the best way to remove stale BSY files. Mystic creates BSY files to ensure another process is not running while node
activity is happening when packets are being moved about, processed etc.

If you are having issues intermittently I'd try and pin down if it is post a fidopoll session with a given node / system connecting to. I can say here I poll internally a BinkD system without issue.

This last busy period lasted the better part of 24 hours. I have since added a Mystic event to delete all busy files once every 12 hours until
a better solution to the busy is sue can be found.

BSY files that remain are symptomatic of a wider issue. I could even be
corrupt echomail node data within your Mystic, I've seen that in the past but if the problems seem to be related to a failed fidopoll then yep that could easily cause an issue if the exe locks up..

--- Mystic BBS v1.12 A43 2019/03/03 (Windows/32)
* Origin: Agency BBS | Dunedin, New Zealand | agency.bbs.nz (3:770/100)

On 18 Sep 2019 at 08:17p, Jeff Smith pondered and said...

I am contining to have a problem with stale or undeleted busy files. I

[snip]

Sep 18 00:42:01 FIDOPOLL Version 1.12 A43 2019/03/02
Sep 18 00:42:01 Looking up address "1:3634/12" in nodelist
Sep 18 00:42:01 Scanning 1:3634/12
Sep 18 00:42:01 Queued 0 files (0 bytes) to 1:3634/12
[...]
Sep 18 00:42:01 R: NUL TIME Wed, 18 Sep 2019 01:42:01 -0400
Sep 18 00:42:01 Info TIME Wed, 18 Sep 2019 01:42:01 -0400
Sep 18 00:42:01 R: NUL VER binkd/1.1a-99/Linux binkp/1.1
Sep 18 00:42:01 Mailer binkd/1.1a-99/Linux binkp/1.1
Sep 18 00:42:01 R: ADR 1:3634/12@fidonet 1:3634/0@fidonet
1:123/0@fidonet 1:18/0@fidonet 1:1/120@fidonet 432:1/139@vkradio

did it hang there Jeff?

was this IPv4 or 6 ?

--- Mystic BBS v1.12 A43 2019/03/03 (Windows/32)
* Origin: Agency BBS | Dunedin, New Zealand | agency.bbs.nz (3:770/100)

Hello Paul!

20 Sep 19 20:11, you wrote to me:

On 18 Sep 2019 at 08:17p, Jeff Smith pondered and said...

I am contining to have a problem with stale or undeleted busy
files. I

[snip]

Sep 18 00:42:01 FIDOPOLL Version 1.12 A43 2019/03/02
Sep 18 00:42:01 Looking up address "1:3634/12" in nodelist
Sep 18 00:42:01 Scanning 1:3634/12
Sep 18 00:42:01 Queued 0 files (0 bytes) to 1:3634/12
[...]
Sep 18 00:42:01 R: NUL TIME Wed, 18 Sep 2019 01:42:01 -0400
Sep 18 00:42:01 Info TIME Wed, 18 Sep 2019 01:42:01 -0400
Sep 18 00:42:01 R: NUL VER binkd/1.1a-99/Linux binkp/1.1
Sep 18 00:42:01 Mailer binkd/1.1a-99/Linux binkp/1.1
Sep 18 00:42:01 R: ADR 1:3634/12@fidonet 1:3634/0@fidonet
1:123/0@fidonet 1:18/0@fidonet 1:1/120@fidonet 432:1/139@vkradio

did it hang there Jeff?

Yup. The above is where the log file ended. I have noticed that
there have been previous and subsequent connections to the same
node that completed succesfully

was this IPv4 or 6 ?

In this instance it was with a IPv4 only specification. Previously
I had Mystic setup to use either IPv specification for that node.


Jeff


--- Mystic v1.12 A43 (2019/03/02) GoldED+/LNX 1.1.5-b20170303
* Origin: The Twisted Balloon - bbs.twisted-balloon.org (1:14/6)

Hello Paul!

I should also state that I have been watching the Mystic system
here more closely the last three days. And I have had to issue
the FIDOPOLL Killbusy All command manually 2-3 times a day to
allow Fidopoll and/or Mutil to run as the configured events to
poll or process would be prevented to run. This is with a event
to run a FIDOPOLL KILLBUSY ALL every 12 hours.

That raises a questiom. If there is an existing stale fidopoll
busy file in the semaphore directory. Will the "Fidopoll Killbusy
All" even be able to run as a Mytic event or will it also be
prevented to run?

An alternative would be to run the Killbusy event as Cron event
apart from Mystic.


Jeff


--- Mystic v1.12 A43 (2019/03/02) GoldED+/LNX 1.1.5-b20170303
* Origin: The Twisted Balloon - bbs.twisted-balloon.org (1:14/6)