Real-Time Backup issues

Seafile Server
1

Hi all,
has anybody same connectivity problems Realtime Backup Server
seafile pro 6.1.3 @ debain
Seafile in the Cluster MariaDB Galera Cluster setup, one node of it is used as replication master for RTB (see https://manual.seafile.com/deploy_pro/real_time_backup.html)
RTB is running well for some libraries, but there are following issues:

  • some libraries cannot be synced:
./seaf-backup-cmd.sh status                                           
Total number of libraries: 68
Number of synchronized libraries: 66
Number of libraries waiting for sync: 0
Number of libraries syncing: 0
Number of libraries failed to sync: 2

List of syncing libraries:

List of libraries failed to sync:
1a2bfe1b-4f2a-4835-af23-fdd42a694bbd
4548a4ed-7163-422d-91c6-22e1d888cd58

and an attempt of forced sync:

sudo -u seafile ./seaf-backup-cmd.sh sync --force 4548a4ed-7163-422d-91c6-22e1d888cd58
Failed to sync repo 4548a4ed-7163-422d-91c6-22e1d888cd58: Failed to get commit ids from primary.

The libraries have been checked on master:

sudo -u seafile ./seaf-fsck.sh 1a2bfe1b-4f2a-4835-af23-fdd42a694bbd

Starting seaf-fsck, please wait ...

[05/09/18 14:10:26] fsck.c(595): Running fsck for repo 1a2bfe1b-4f2a-4835-af23-fdd42a694bbd.
[05/09/18 14:10:26] fsck.c(422): Checking file system integrity of repo 170323(1a2bfe1b)...
[05/09/18 14:10:26] fsck.c(659): Fsck finished for repo 1a2bfe1b.

seaf-fsck run done

Done.

Everything seems to be OK…

  • connectivity problems on RTB Server :
    /var/log/mysql/error.log
2018-05-09 10:55:45 140618125444864 [Warning] Aborted connection 35 to db: 'seafile-db' user: 'seafile' host: 'localhost' (Got an error reading communication packets)
2018-05-09 10:55:45 140618125043456 [Warning] Aborted connection 34 to db: 'seafile-db' user: 'seafile' host: 'localhost' (Got an error reading communication packets)
2018-05-09 10:55:54 140617968527104 [Warning] Aborted connection 165 to db: 'seahub-db' user: 'seafile' host: 'localhost' (Got an error reading communication packets)
2018-05-09 10:55:54 140617968928512 [Warning] Aborted connection 118 to db: 'seahub-db' user: 'seafile' host: 'localhost' (Got an error reading communication packets)
2018-05-09 10:55:54 140618125244160 [Warning] Aborted connection 83 to db: 'seahub-db' user: 'seafile' host: 'localhost' (Got an error reading communication packets)
2018-05-09 10:56:06 140617969129216 [Warning] Aborted connection 556 to db: 'seafile-db' user: 'seafile' host: 'localhost' (Got an error reading communication packets)
2018-05-09 10:56:06 140617967924992 [Warning] Aborted connection 553 to db: 'seafile-db' user: 'seafile' host: 'localhost' (Got an error reading communication packets)

A full access for seafile user to both tables is granted on the host, it is possible to login into mysql client via mysql -useafile -hlocalhost -p. Any idea?

2

Hi,

I have similar warnings. I had warnings like these with an another service too (z-push).
The problem was that z-push uses persistent DB connection, and the MySQL server closes the connection (this is the warning we have) after 10s because the connection is open, but inactive. I “solved” my problem (z-push): disabled persistent DB connections. I think the problem is same with seahub. And an another idea to fix: we could increase the connection timeout value in DB’s configs.

It would be great if a developer could give us some information about Seafile’s DB usage.

Norbert

3

I have exactly the same warnings/errors. I tried increasing timeouts on mysql but it made no difference. Would love to hear the solution if anyone has found one

3 Likes