s2s overflow and s2s_syncer errors

[igardais]

Hi,

I have multiple issues when doing a RS2211RP+ to RS2211RP+ backup.
Both runs the same revision of the firmware :
- Linux SYNO-1 2.6.32.12 #1955 SMP Sat Nov 26 14:52:27 CST 2011 x86_64 GNU/Linux synology_x86_rs2211rp+
- Linux syno-2 2.6.32.12 #1955 SMP Sat Nov 26 14:52:27 CST 2011 x86_64 GNU/Linux synology_x86_rs2211rp+


First, there is a *lot* of overflow messages in the /var/log/message. Always the same message separated by few seconds.
Jan 10 05:21:31 s2s_monitor[14412]: (s2s_utils.c:857) overflow occurs

Second, rsync failed on some sync.
From the source synology, the messages are :
Jan 10 11:53:11 s2s_syncer: sync share:[/volume1/it-pra_vm] failed. ret=[43]
Jan 10 11:53:12 s2s_syncer: Failed to execute rsync synocmd at updateTaskProcess. source=[/tmp/filecFWi2m], target=[admin@172.31.48.12::], passwd:[/tmp/s2sPass_it-pra_vm] ret=44

On the destination, the messages are :
Jan 10 11:47:43 rsyncd[15277]: io timeout after 600 seconds when processing (internal/print/print-flat.vmdk) -- exiting
Jan 10 11:47:43 rsyncd[15272]: rsync error: timeout in data send/receive (code 30) at io.c(254) [receiver=3.0.4]
Jan 10 11:47:43 rsyncd[15272]: rsync error: rsync service is no running (code 43) at io.c(682) [generator=3.0.4]

Both Synologies are time-synced correctly.
The file internal/print/print-flat.vmdk size is 60GB.

Note that the VMs are not in use : this is a copy of a backup. (Yeah, I know).


Then at last there is this message, which seems to be related to the first one :
Jan 10 12:19:42 s2s_syncer: sync share:[/volume1/it-db] failed. ret=[24]


Any clues on this ?

Thanks,
Ionel

[bwestover]

I'm getting this same error periodically in my logs:

s2s_syncer: Failed to execute rsync synocmd at updateTaskProcess. source=[/tmp/<somefile>], target=[admin@<mytargetserver>::], passwd:[/tmp/s2sPass_Sync <Sharename>] ret=44

In my case however, it is the Shared Folder Sync option that seems to be the culprit. Some googling turned up that "ret=44" means "out of space". In fact I am also seeing this error in my logs: Shared Folder Sync failed due to watch share [sharename] failed. Num of subfolders in the share exceeds the maximum.(error:[No space left on device])

I have been made aware of a 100K folder limit, which I am under. I do however have alot of files (over 1 million), and some that are fairly large (several gigabytes). I don't store VMDKs on this volume though, so nothing as big as 60gb like your seeing.

Still, I wonder if the upper threshold of what can be cached is what we're crashing into?

Have you contacted support?

[bwestover]

Just a follow up on this. It turned out that my second error "No space left on device" was in fact due to me being over the 102400 folder limit.

It was a bit tricky, the folder count you see as exposed via SAMBA is not the same as the actual folders on the disk. There are hidden folders called @eaDir that you can only see at the command line of the device.

In my case those folders (about 19K of them) were pushing me from being under the limit, to being well over.

See this thread for more details: viewtopic.php?f=160&t=45110

[igardais]

Sorry for not giving feedback earlier.
I'll check for the 100k limit but as this only a VM datastore, there is not a huge level of directory imbrication.
Thus still seeing the error, even with DSM 4.0

[bwestover]

So just to clarify, I was getting the error "Failed to execute rsync synocmd...ret=[44]" but I am nearly certain that was due to running out of inotify_watches.

Now that I have adjusted my share to be under the limit, I never see that one anymore. The other error though "s2s_syncer: sync file:[<path>] failed. ret=[24]" I still get periodically, and I have no idea why.

See this thread for more information on that.

viewtopic.php?f=160&t=36447

[igardais]

On my company synology, s2s_syncer is still making a looooot of error messages :


Apr 13 03:21:02 s2s_monitor[10422]: (s2s_utils.c:857) overflow occurs

There are 400 entries for today /var/log/messages in 8 hours :

SYNO-1> grep -c "overflow occurs" /var/log/messages
400

Could you please correct this bug ?
Is there a bugzilla somewhere ?

Thanks,
Ionel