You are here
Jacob - Wed, 2014/07/02 - 19:02
Cut & Paste from email with Liraz:
--------------------------------------
There's a trick you can use to throttle backup/restore speeds:
http://www.turnkeylinux.org/forum/support/20120327/throttle-backup-and-restore-bandwidth-speeds-tklbam
To debug what is going on at the network level I recommend trying out
nload, iftop or iptraf.
To stop the backup and try resuming it, just Ctrl-C and then rerun
whatever command you used - using exactly the same arguments.
Let me know how it goes.
You're welcome,
Liraz
PS: we should post your question and my answer to the forum so that
anyone that runs into the same problem (or has found a solution) can
chip in.
On 02/07/14 17:51, Jacob de Greef wrote:
Hi Liraz,
Sorry to bother you via mail. I searched the forums, but could not find any solution.
I installed the new version of the fileserver app, based on deb 7. Restored my old files. Started a fresh TKLBAM backup. This backup has been running for over 40 hrs (approx. 5 Gigabyte). My upload is normally between 1 and 3 Mbit/sec, so it should have finished by now. I cannot see any progress or errors, but I notice my entire bandwidth is clogged, even simple web browsing takes ages. On the Hub dashboard I see: 'Backup in progress… 0 bytes'
What can I do to check for status / progress / errors?
How can I safely pause / throttle the backup process and restart it at night?
Any other advice?
Thanks,
Jacob
Forum:
Tags:
trickle only for new processes(?)
I found out about trickle in the forums, but thought it only affects new started processes and will not in real time affect already running jobs.
So I gues I will Ctrl-C the currenct process, then install and run trickle, and then try to resume the backup.
One more question, what happens when after 24 hrs TKLBAM tries to initiated the incremental backup while the initial backup has not finished?
TKLBAM tries to pick up from
TKLBAM tries to pick up from where it left. It doesn't matter how much time has passed. Some things such as databases are only serialized (mysqldump'ed) once and are not reserialized when you resume. If new files have changed in between it won't pick up on that until the next backup.
Thanks Jacob. FWIW, in case
Thanks Jacob. FWIW, in case anyone finds it weird that we're pasting emails into the forum - the idea is to make it easier for others that run into a similar problem (or know the solution) to benefit or chip in.
log files
Tried to post the tklbam-backup logfile here. But the mod system regards it as spam :( Too many s3 hyperlinks I guess. Any way to get aroud that?
Try attaching them to your OP
You can only do that with the first post of a thread, so you'll need to edit that. But that shouldn't be an issue for the spam filter (although following the tweaks that Liraz did recently; when logged in you should be able to avoid the spam filter altogether now... Not sure why that isn't working...?!)
There was a bug, I think I squashed it
exerpt log file
Trying to copy-paste a snippet of /var/log/tklbam-backup
I have a mobile broadband connection. The ISP is NATting the connection, so I don't have a public IP address. The quality of the connection varies :( sometimes it seems to drop and reconnects after a couple of minutes. I cannot get a decent log out of the MiFi device, and Huawei doesn't give information on how to put it in bridge mode.
Anyway, here's the snippet. Please shoot if you see any indication what may cause the interupted backups.
TKLBAM is Liraz's baby and I'm not really much of a programmer
So his perspective will no doubt be much more enlightening than mine...
From my casual glance through your log (plus the other info that you have mentioned) it seems pretty clear to me that it is a connection issue. One thing that you may be able to try is reducing the volume size. Smaller volumes may have a better chance of getting through without being timed out? Probably worth a try...
[Edit]: Oops I just realised that it looks like you have already tweaked the volume size down to 5MB (from the default of 50MB).
Yes, this looks like a connection issue
Sometimes the Hub can be slow to update
Although it is strange that the Sun & Mon incrementals aren't there...
Have you checked your TKLBAM logs (on your server) to see what they say? I.e. check that the backups ran and what if any errors may have been raised. IIRC log should be /var/log/tklbam (or similar).
You could use 'tklbam-list' from your server to double check the Hub's backup records. It shouldn't make any difference as the 'tklbam-list' command gets it's info from the same dataset as the Hub UI but perhaps it's worth a try anyway...?!
If you can confirm that the backup definitely ran successfully (from your logs - copy/paste the relevant bit into a new comment) and that it doesn't show up in the Hub UI (perhaps copy/paste from hub-list) then we should lodge a pretty convincing bug report (to the TKL Issue Tracker).
Add new comment