On 2024-02-13 02:49, Eric Bollengier wrote:
Hello Peter,

Without the ReadFifo directive, it's unlikely to cause a problem,
Unlikely maybe but that is the problem and I can even reproduce it!
My setup is based on ubuntu 22.04 LTS (was trying debian but align is broken there) using the community repo

  debhttps://www.bacula.org/packages/<redacted>/debs/13.0.3 jammy main

and the file daemon output is pretty clear, we are not at this file.

The file daemon output shows last file that worked, not the file it is trying to backup.

To reproduce I did

*1 - create a fileset that backups just /tmp/debug*

FileSet {
    Name = "debugfs2"
    Ignore FileSet Changes = yes
    Include {
      Options {
        signature=MD5
        honor nodump flag=yes
        noatime=yes
        keepatime = no
        sparse=yes
        exclude = yes
        wild = *~
        wild = *.tmp
        }
      File = "/tmp/debug"
      }
    }

*2 - create a pipe ("mkfifo random_pipe") and a plane file ("date >a_red_herring") in /tmp/debug*

peters@quark:/tmp/debug$ find /tmp/debug/ -ls
    11017      0 drwxr-xr-x   2 peters   peters         80 Feb 13 08:47 
/tmp/debug/
    11031      4 -rw-r--r--   1 peters   peters         32 Feb 13 08:47 
/tmp/debug/a_red_herring
    11029      0 prw-r--r--   1 peters   peters          0 Feb 13 08:46 
/tmp/debug/random_pipe
peters@quark:/tmp/debug$

*3 - start a backup;
*

root@quark:~#echo run BackupQ_quark FileSet="debugfs2" Level=Full yes|bconsole

*4 - confirm it hangs*

root@quark:~# echo stat client=quark-fd|bconsole #CLIENTSTAT
Connecting to Director quark:9101
1000 OK: 10002 techwiz-dir Version: 13.0.3 (02 May 2023)
Enter a period to cancel a command.
stat client=quark-fd
Connecting to Client quark-fd at quark:9102

quark-fd Version: 13.0.3 (02 May 2023)  x86_64-pc-linux-gnu-bacula-enterprise 
ubuntu 22.04
Daemon started 12-Feb-24 23:55. Jobs: run=5 running=1.
 Heap: heap=856,064 smbytes=603,907 max_bytes=1,219,047 bufs=178 max_bufs=429
 Sizes: boffset_t=8 size_t=8 debug=0 trace=0 mode=0,0 bwlimit=0kB/s
 Crypto: fips=N/A crypto=OpenSSL 3.0.2 15 Mar 2022
 Plugin: bpipe-fd.so(2)

Running Jobs:
JobId 315 Job BackupQ_quark.2024-02-13_08.48.07_46 is running.
    Full Backup Job started: 13-Feb-24 08:48
    Files=1 Bytes=40 AveBytes/sec=0 LastBytes/sec=2 Errors=0
    Bwlimit=0 ReadBytes=32
    Files: Examined=1 Backed up=1
    Processing file: /tmp/debug/a_red_herring
    SDReadSeqNo=8 fd=5 SDtls=1
Director connected using TLS at: 13-Feb-24 08:55
====

*5 - release the job by sending something to the pipe; *

root@quark:~# echo >/tmp/debug/random_pipe

*6 - confirm the job finished*

root@quark:~# echo 'llist jobid=315'|bconsole
Connecting to Director quark:9101
1000 OK: 10002 techwiz-dir Version: 13.0.3 (02 May 2023)
Enter a period to cancel a command.
llist jobid=315
Automatically selected Catalog: MyCatalog
Using Catalog "MyCatalog"
           JobId: 315
             Job: BackupQ_quark.2024-02-13_08.48.07_46
            Name: BackupQ_quark
     PurgedFiles: 0
            Type: B
           Level: F
        ClientId: 32
      ClientName: quark-fd
       JobStatus: T
       SchedTime: 2024-02-13 08:48:07
       StartTime: 2024-02-13 08:48:09
         EndTime: 2024-02-13 08:56:40
     RealEndTime: 2024-02-13 08:56:40
        JobTDate: 1,707,832,600
    VolSessionId: 66
  VolSessionTime: 1,707,779,559
        JobFiles: 3
        JobBytes: 40
       ReadBytes: 32
       JobErrors: 0
 JobMissingFiles: 0
          PoolId: 2
        PoolName: File-Full
      PriorJobId: 0
        PriorJob:
       FileSetId: 7
         FileSet: debugfs2
        HasCache: 0
         Comment:
        Reviewed: 0

You have messages.
root@quark:~#

/ps


The problem can be somewhere else, and a good start is a "status dir"
and "status storage".

Best Regards,
Eric

On 2/13/24 06:23, Peter Sjoberg wrote:

Actually, I think I found the root cause - a pipe!

The file listed in the client status is not the problem but close to it is a pipe (maybe next file) and that is what causing the issue in all cases.
I striped down the directory to just one file and it still fails

root@defiant1:/home/debug# find .zoom -ls
  23855106      4 drwx------   4 sys      adm          4096 Feb 13 00:11 .zoom   23855107      4 drwxrwxr-x   2 ba       ba           4096 Feb 13 00:11 .zoom/data   23855116      0 prw-r--r--   1 ba       ba              0 Mar 12  2021 .zoom/data/com.zoom.ipc.confapp__res   23855110      4 drwxrwxr-x   2 sys      adm          4096 May 5  2020 .zoom/reports
root@defiant1:/home/debug#


and if I send something to the pipe the job finish ok and restoring the job did include the pipe.

/ps

On 2024-02-12 23:36, Peter Sjoberg wrote:

In short - no and no, no special files
Also, while I haven't waited forever I have left it for several hours so it's not like it's just some house cleaning left.
It does happen on different servers, sample below is my laptop.

=============== My fileset:
FileSet {
     Name = "debugfs2"
     Ignore FileSet Changes = yes
     Include {
       Options {
         signature=MD5
         honor nodump flag=yes
         noatime=yes
         keepatime = no
         sparse=yes
         exclude = yes
         wild = *~
         wild = *.tmp
         }
       File = "/home/ba/.zoom"
       }
     }

============= run command
echo run BackupQ_defiant1 FileSet="debugfs2" Level=Full yes|bconsole


============= client status when hung
root@quark:~# echo stat client=defiant1-fd|bconsole #CLIENTSTAT
Connecting to Director quark:9101
1000 OK: 10002 techwiz-dir Version: 13.0.3 (02 May 2023)
Enter a period to cancel a command.
stat client=defiant1-fd
Connecting to Client defiant1-fd at defiant1:9102

defiant1-fd Version: 13.0.3 (02 May 2023) x86_64-pc-linux-gnu-bacula-enterprise ubuntu 22.04
Daemon started 12-Feb-24 23:22. Jobs: run=2 running=1.
 Heap: heap=856,064 smbytes=606,583 max_bytes=794,675 bufs=188 max_bufs=203
 Sizes: boffset_t=8 size_t=8 debug=0 trace=0 mode=0,0 bwlimit=0kB/s
 Crypto: fips=N/A crypto=OpenSSL 3.0.2 15 Mar 2022
 Plugin: bpipe-fd.so(2)

Running Jobs:
JobId 255 Job BackupQ_defiant1.2024-02-12_23.29.48_18 is running.
    Full Backup Job started: 12-Feb-24 23:29
    Files=2 Bytes=0 AveBytes/sec=0 LastBytes/sec=0 Errors=0
    Bwlimit=0 ReadBytes=0
    Files: Examined=2 Backed up=2
    Processing file: /home/ba/.zoom/logs
    SDReadSeqNo=8 fd=5 SDtls=1
Director connected using TLS at: 12-Feb-24 23:30
====


================ Content of that directory

root@defiant1:~# find  /home/ba/.zoom -ls
  15728688      4 drwx------   7 ba       ba           4096 Apr 28  2021 /home/ba/.zoom   16130303      4 drwx------   2 ba       ba           4096 Apr 28  2021 /home/ba/.zoom/screenCapture   16130301      4 drwxrwxr-x   2 ba       ba           4096 Feb 12 18:47 /home/ba/.zoom/logs   16023538      4 drwxrwxr-x   4 ba       ba           4096 Apr 28  2021 /home/ba/.zoom/data   16023540      0 prw-r--r--   1 ba       ba              0 May  5  2020 /home/ba/.zoom/data/com.zoom.ipc.assistantapp__res   16023541      0 prw-r--r--   1 ba       ba              0 Mar 12  2021 /home/ba/.zoom/data/com.zoom.ipc.confapp__req   16023539      0 prw-r--r--   1 ba       ba              0 Apr 28  2021 /home/ba/.zoom/data/com.zoom.ipc.assistantapp__req   16130305      4 drwx------   2 ba       ba           4096 Mar 23  2021 /home/ba/.zoom/data/VirtualBkgnd_Custom   16131475   1564 -rw-------   1 ba       ba        1597940 Mar 23  2021 /home/ba/.zoom/data/VirtualBkgnd_Custom/{ff6d8a57-d810-4dd2-bf1b-8366c063728f}   16023542      0 prw-r--r--   1 ba       ba              0 Mar 12  2021 /home/ba/.zoom/data/com.zoom.ipc.confapp__res   16023545     52 -rw-------   1 ba       ba          53248 Apr 28  2021 /home/ba/.zoom/data/zoomus.enc.db   16130304      4 drwx------   2 ba       ba           4096 Mar 23  2021 /home/ba/.zoom/data/ConfAvatar   16131472     36 -rw-------   1 ba       ba          36397 Mar 23  2021 /home/ba/.zoom/data/ConfAvatar/conf_avatar_6c72761c1ad5cc6f485dce3966cbb705_100   16131473      4 -rw-------   1 ba       ba           1020 Mar 23  2021 /home/ba/.zoom/data/ConfAvatar/conf_avatar_9e6b3f01c5d33a2052c2681a42b4e659_100   16131474      4 -rw-------   1 ba       ba           1020 Mar 23  2021 /home/ba/.zoom/data/ConfAvatar/conf_avatar_e977cbed2632f5b11882e92e31f32516_100   16023544      8 -rw-------   1 ba       ba           5120 Mar 23  2021 /home/ba/.zoom/data/zoommeeting.enc.db   16130302      4 drwxrwxr-x   2 ba       ba           4096 May  5  2020 /home/ba/.zoom/reports   16130300      4 drwx------   2 ba       ba           4096 Apr 28  2021 /home/ba/.zoom/im
root@defiant1:~#

/ps


On 2024-02-12 20:03, Gary R. Schmidt wrote:
On 13/02/2024 11:08, Phil Stracchino wrote:
On 2/12/24 18:35, Peter Sjoberg wrote:
Hi all

I have a strange problem and (on my system) reproducible problem. When I do backup of some directories then bacula-fd just hangs and never complete. The directories in question are not very strange and backup of them works find with older versions of -fd


Silly question:  Do the problem directories contain named pipes or sockets?

Another possibly silly question: Are there any soft links that may cause a loop?

    Cheers,
        Gary    B-)


_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users




_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users




_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users

Reply via email to