[Resolved] Issue with Linux agent (version 11)

Forgot to say....

If I try and telnet to the IP of the box on port 9876 I can see that there is something accepting connects there, so I know that there's nothing getting bounced etc by firewalls etc.

Ol

So a bit more progress here...

I've managed to get the management server to add the agent again, it did exactly the same thing - added it, all seemed well and then suddenly it disconnected.

I noticed that the MMS service died around this time, restarting it hasn't helped.

Are there any debug files I can post? to try and sort this

At the moment i'm going to start looking at other solutions as we need something that works quickly and easily.

Hi Oli!

Thank you for bringing this issue to our attention!

Looks like the issue is caused by mms service crashing. Usually it happens when snapapi is incorrectly build, though cannot say for sure without additional logs. Can you send me, please: 

1. Linux System Report
2. Logs from /var/lib/acronis
3. Strace log: 
   GUI
   - First of all install strace if you don't have it (for RedHat the command is)
   # yum install strace
   - Rename the backup and recovery daemon:
   # mv /usr/lib/Acronis/BackupAndRecoveryAgent/BackupAndRecoveryService /usr/lib/Acronis/BackupAndRecoveryAgent/Agent
   - Launch the console:
   # acronis_console
   and create a new task. The task should be scheduled to a manual start. You can find the script of the newly created task in the /var/lib/Acronis/TrueImage/Scripts/ directory (choose the latest if there are several scripts).
   - Launch the daemon under the strace:
   # export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/lib/Acronis/system_libs:/usr/lib/Acronis/Common:/usr/lib/Acronis/BackupAndRecovery/Common
   # /usr/bin/strace -fF -s 1024 -o /root/strace.log /usr/lib/Acronis/BackupAndRecoveryAgent/Agent -script:/var/lib/Acronis/TrueImage/Scripts/[ScriptName].tmp.tib.tis

   Like:
   # /usr/bin/strace -fF -s 1024 -o /root/strace.log /usr/lib/Acronis/BackupAndRecoveryAgent/Agent -script:/var/lib/Acronis/TrueImage/Scripts/AAF7A67A-32B5-C4F5-F846-3B0F1BB2CD85.tmp.tib.tis
   - Reproduce the issue
   - You will find the log here: /root/strace.log
   - Rename the daemon back:
   # mv /usr/lib/Acronis/BackupAndRecoveryAgent/Agent /usr/lib/Acronis/BackupAndRecoveryAgent/BackupAndRecoveryService
   CLI
   Use the following command to get the strace output: 
   # /usr/bin/strace -fF -ttt -s 1024 -o /root/strace.log /usr/sbin/trueimagecmd [options]
   or
   # /usr/bin/strace -fF -ttt -s 1024 -o /root/strace.log /usr/sbin/trueimagemnt [options]
   It should look like
   # /usr/bin/strace -s 1024 -fF -ttt -o /root/cli.log /usr/sbin/trueimagecmd --create --partition:1-1 --filename:/backup.tib

Send these logs to me, this data should be sufficient for localizing cause. If you have any questions, let me know, I will be glad to help you.

Thank you.

we have exactly the same problem you made it farther then we did at this time we are unable to install the client Red Hat 5.5
[Acronis 01290660] MMS service not stopping.

Hi data center,

Thanks for the case number and details.

I cannot confirm that Oli actually has the same issue that you do, your's a specific one. It's related to the RedHat kernel bug actually. Long story short - virtio blk devices do not support scsi ioctls. Our developers managed to workaround this conflict: no ioclts are sent on the machine where virtio is detected, the fix will be available in the upcoming update. 

I understand how urgent the situation is, and will contact Development team, maybe they will be able to provide us with a workaround or a temporary solution. I will contact you via PM once I get any news from them.

Thank you.

Hi Yana,

Sorry for the delay - I've been away with the festive break etc.

Where should I send these details to?

Thanks

Ol

Hello Yana,

I have the the same problem as user "data center". MMS process on CentOS 5 host freeze when I'm trying to add machine to AMS. In dmesg is following text:

INFO: task mms:5663 blocked for more than 120 seconds.
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
mms D ffff81000253eaa0 0 5663 1 5688 5648 (NOTLB)
ffff81004bb997d8 0000000000000082 0000000000000220 ffff81007fcfdc40
ffff81007f2a6058 0000000000000005 ffff81007fb137a0 ffff81007ffb9100
0004071bb9b5dc5d 000000000000d992 ffff81007fb13988 00000001801478fc
Call Trace:
[] :virtio_pci:vp_notify+0x0/0x1c
[] wait_for_completion+0x79/0xa2
[] default_wake_function+0x0/0xe
[] blk_execute_rq+0x98/0xc0
[] blk_recount_segments+0x17/0x28
[] sg_io+0x258/0x356
[] scsi_cmd_ioctl+0x1d2/0x3b5
[] cfq_prio_tree_add+0x9b/0xac
[] enqueue_task+0x41/0x56
[] __activate_task+0x56/0x6d
[] blkdev_driver_ioctl+0x5d/0x72
[] :dm_mod:linear_ioctl+0x7c/0x8e
[] do_lookup+0x65/0x1e6
[] :dm_mod:dm_blk_ioctl+0x99/0xbe
[] blkdev_driver_ioctl+0x5d/0x72
[] blkdev_ioctl+0x63c/0x697
[] mutex_lock+0xd/0x1d
[] :dm_mod:dm_blk_open+0x39/0x5e
[] do_open+0x252/0x311
[] blkdev_open+0x0/0x4f
[] cp_new_stat+0xe5/0xfd
[] block_ioctl+0x1b/0x1f
[] do_ioctl+0x21/0x6b
[] vfs_ioctl+0x457/0x4b9
[] sys_newfstat+0x20/0x29
[] sys_ioctl+0x59/0x78
[] system_call+0x7e/0x83

I made some tests and it looks like problem with virtio blk scsi ioctls. It's any progress from devel team to solve this problem? Exists any workarround for centos 5 machines running on KVM with virtio block device?

I'm trying to solve it via local (Czech) support team (They provide me acronis support case number: 01463990).

Thank you
Milan

This issue is resolved with ABR build 17437 released 21-March-2012.