I upgraded from rc19 to rc20 and hit a problem, there is now a 5 minute delay during array start.

 

I can't really tell much from the logs, other than seeing a 5 minute gap between anything interesting:

Jan  6 18:42:28 Tower emhttpd: shcmd (120): /usr/local/sbin/update_cron
Jan  6 18:45:25 Tower nginx: 2018/01/06 18:45:25 [error] 8552#8552: *1 upstream timed out (110: Connection timed out) while reading upstream, client: 192.168.10.138, server: , request: "POST /update.htm HTTP/2.0", upstream: "http://unix:/var/run/emhttpd.socket:/update.htm", host: "<myhash>.unraid.net:2443", referrer: "https://<myhash>.unraid.net:2443/Main"
Jan  6 18:47:29 Tower root: Fix Common Problems Version 2018.01.06

I don't think this was an issue in rc19, but it is possible I wasn't looking closely.

 

Note that if I boot in safe mode or start the array in maintenance mode, there is no delay all when starting the array.  Does it *have* to be a plugin issue or are there other things that happen behind the scenes?

 

For the attached diagnostics, I have already:

• disabled array autostart
• disabled dockers and vms
• unplugged the flash drive for my unraid VM (since it is disabled anyway)
• unplugged my USB card reader
• disabled (but not uninstalled) the cache_dirs, auto turbo, ransomware, recyclebin, and file integrity plugins

 

Does anyone see any possible reason for this or do I need to actually start uninstalling plugins?  Is anyone else seeing these delays?

 

Oh - one of my disks is nearly full (XFS, about 100GB free out of 4TB). That wouldn't be the problem would it?

 

 

 

BTW, there is also a privacy issue in the diagnostics - when nginx times out, we need to suppress the url that timed out (I did this manually in the attached):

Jan  6 18:45:25 Tower nginx: 2018/01/06 18:45:25 [error] 8552#8552: *1 upstream timed out (110: Connection timed out) while reading upstream, client: 192.168.10.138, server: , request: "POST /update.htm HTTP/2.0", upstream: "http://unix:/var/run/emhttpd.socket:/update.htm", host: "<myhash>.unraid.net:2443", referrer: "https://<myhash>.unraid.net:2443/Main"

 

tower-diagnostics-20180106-1847.zip

Edited January 9, 20188 yr by ljm42

1 hour ago, ljm42 said:

5 minute

If you grabbed the update for FCP earlier today, then there is a mistake in it that will potentially cause a 5 minute delay on array start.  2018.01.06a fixes that mistake

Thank you @Squid, that was it  

 

On the good news side, it provided an opportunity to notice the privacy issue in the diagnostics  I won't mark this solved yet, to give @limetech or @bonienl the chance to see that issue.

Thanks for fixing this privacy issue @bonienl