About 3 hours ago I started seeing "Timed out trying to read data from the socket stream!" in the MXdiags log. A lot of them. CMX (3.9.1) had been running for about 3 days and this seemed to have started at 9AM here - on the hour. As far as I know nothing on the Pi or with my network changed recently. I chatted with my hosting provider to see if they had any outages or changed any policies - but they checked logs and said everything looked find on their side. I ran FTP from the command line and didn't seem to have any issues. Also, on the hosted server I could see files were getting updated - I run "watch ls -alt --full-time", and I can see the files updating.
I turned on FTP logging and could spot any obvious issues - though I'm not terribly familiar and there is a lot of output. I was looking for a correlation between the 2 logs by doing a "tail -f" on each side-by-side. What I noticed was when I saw the message as issue in the MXdiags log, I seemed to see 3 messages in the FTP log:
Status: Disposing FtpSocketStream...
Status: Disposing FtpSocketStream...
Status: Disposing FtpSocketStream...
That seemed to be the only time I saw those 3 repeated lines in the FTP log, i.e. when "Timed out trying to read data from the socket stream!" occurred in the MXdiags log.
Scanning the MXdiags log I also noticed a periodic instance of this message:
Warning, previous web update is still in progress, first chance, skipping this interval
That seems suspiciously related to the 2 minute intervals on updates - but this isn't consistent.
viewtopic.php?p=134421 caught my eye when searching, but that seemed to be the result of a firewall issue. There's certainly nothing I did explicitly with firewalling this morning.
Appreciate any thoughts-
Might be of interest how I added a timestamp prefix to the ftp log...
Code: Select all
tail -f /home/pi/CumulusMX/ftplog.txt | while read LINE ; do echo "`date +'%T.%N '` ${LINE}" ; done | tee -a ftplog_a.txt