Project

General

Profile

Bug #2026

Greyhole starts faster then MariaDB - Critical Error in Log "Can't connect to database"

Added by micmas almost 5 years ago. Updated almost 5 years ago.

Status:
Closed
Priority:
Normal
Assignee:
Category:
Greyhole
Target version:
Start date:
04/07/2016
Due date:
% Done:

0%


Description

The F23-system is installed new from scratch - you find it at
http://paste.fedoraproject.org/349268/45976696/ (bug 2222)

Every system start I find this errors in the Greyhole log: ###
Apr 07 12:49:25 CRITICAL initialize: Can't connect to database: SQLSTATE[HY000] [2002] No such file or directory
Apr 07 12:49:26 CRITICAL initialize: Can't connect to database: SQLSTATE[HY000] [2002] No such file or directory
Apr 07 12:49:27 CRITICAL initialize: Can't connect to database: SQLSTATE[HY000] [2002] No such file or directory
Apr 07 12:49:28 CRITICAL initialize: Can't connect to database: SQLSTATE[HY000] [2002] No such file or directory
Apr 07 12:50:32 INFO daemon: Greyhole (version 0.9.64) daemon started.
Apr 07 12:50:32 INFO daemon: Optimizing MySQL tables... ###

After that Greyhoel runs fine all the day.

The MariaDB-log shows at the end of the startup:
"12:49:+29+ /usr/libexec/mysqld: ready for connections"

There is a difference of 4 seconds between both. So I think, Greyhole starts to fast. How can I give Greyhole a delay of 5 seconds before starting?
Or it should start after MariaDB is already to connect.

History

#1 Updated by bigfoot65 almost 5 years ago

  • Assignee changed from bigfoot65 to cpg

#2 Updated by cpg almost 5 years ago

  • Status changed from New to Feedback
  • Assignee changed from cpg to micmas

correct. the service file had a dependency on mysqld. i updated it to be mariadb. you can try it with this fix, which also upgrades you to the latest greyhole:

rpm -Uvh http://alpha.amahi.org/tmp/amahi-greyhole-0.9.66-1.x86_64.rpm

if it works well, please close the bug. If not, give us feedback, set it to assigned and assign it back to me. thanks!

#3 Updated by micmas almost 5 years ago

  • Status changed from Feedback to Closed

It runs, no more critical error, only a warning about Samba: ###
Apr 08 11:15:50 INFO daemon: Greyhole (version 0.9.66) daemon started.
Apr 08 11:15:50 INFO daemon: Optimizing MySQL tables...
Apr 08 11:15:50 WARN daemon: Greyhole VFS module for Samba was missing, or the wrong version for your Samba. It will now be replaced with a symlink to /usr/lib64/greyhole/greyhole-samba43.so
Apr 08 11:15:50 INFO daemon: The Samba daemon will now restart...
Apr 08 11:17:20 INFO daemon: Greyhole (version 0.9.66) daemon started.
Apr 08 11:17:20 INFO daemon: Optimizing MySQL tables... ###
After a second reboot the log is clean: ###
Apr 08 11:20:59 INFO daemon: Greyhole (version 0.9.66) daemon started.
Apr 08 11:20:59 INFO daemon: Optimizing MySQL tables... ###

#4 Updated by cpg almost 5 years ago

wait, what is your samba version?

rpm -q samba

(it should not produce that warning..)

#5 Updated by micmas almost 5 years ago

samba-4.3.6.0.fc23.x86_64

During login as root on the server I get a new error message:
"CIFS VFS: Server 127.0.0.1 has not responed in 120 sec. Reconneting ..."

#6 Updated by cpg almost 5 years ago

i'm a little confused about the warning.

upon installation of the RPM, the system checks and it creates a link to the proper vfs module for your samba version (which at this time is 4.3 for fedora 23).

In my case, it created this symlink.

lrwxrwxrwx 1 root root 39 Apr  8 12:12 /usr/lib64/samba/vfs/greyhole.so -> /usr/lib64/greyhole/greyhole-samba43.so

BUT EVEN THEN ... if it's not there, GH creates it.

do you still get this message on subsequente reboots? (or restarts of the GH service)

#7 Updated by micmas almost 5 years ago

The symlinks are o.k.
Yor are right: After rebooting I don't get any messages. Everything runs good.
Thank you.

Also available in: Atom