I have a ReadyNAS running ReadyOS 6.9.1 that is dedicated as a backup server to back up a couple of other units via Rsync. These backups are done every hour, 24/7. This has been working fine for some time.
I decided to add a backup to external disk of the shares on my backup server that contain the backups from my other units. However, the first time I tried this, the machine appeared to crash. No communication, etc. I power cycled it and everything seems to be ok, but the log entries basically stop right before the time the backup to disk would have been done.
I can't think of a way to prove it, but I suspect the problem is that the Rsync backups *to* my backup server are running at the same time as the backup to disk *on* my backup server, i.e clients are writing to shares that are being read at the same time, and there is a conflict. Unfortunately there appears to be no way to schedule backups at any time other than at 5 minutes past the hour. So the only way I know of to avoid this situation is to introduce a gap into my schedule for backing up my clients while the backup server is backing up to disk. This wouldn't be a huge problem, but I'd rather not do that if I can avoid it.
So, I was wondering if there is a way to schedule backups on a more flexible schedule, or some other trick to avoid any potential conflict.
Is your local backup set up with a local source and a remote destination? Or is it the other way around?
If you have a local source, the NAS should be making a snapshot of the source, and then backing up the snapshot. The snapshot is deleted when the backup completes. That should avoid your issue.
Hmm, I'm not sure that the backup of multiple shares to disk at the same time is working. I tried to start one after the other from the web interface and get an indication that three jobs are in progress. I thought they would be queued. I also am getting a msg on the front panel that says "mb_cache_entry_insert+54". And the backups are taking an inordinate amount of time.
BTW I should mention that this is on a legacy Pro 4 machine. So maybe there is not perfect support for the hardware? At any rate, I think I am on my way to another crash. I just rebooted.
That's a good question. I looked and free tells me I have ~460MB free, at startup anyway. Right now the web interface is apparently stuck in the "canceling..." state for one of the backup to disk jobs. This is after a complete power cycle. Do you happen to know how to clear out this state from ssh?
BTW the disk is formatted with ext4. And free is now telling me that free memory +/- buffers and cache I have 248MB. So maybe the memory issue is a possibility. But I've been running this device w/ OS6 since like september w/ 0 problems.
Thanks for the idea to upgrade memory. Do you happen to know the specs for ram expansion on the pro 4, or how to find it out?
Irrespective of the RAM issue, I think there is something the device does not like about doing backups to my external disk. I can get the device to crash just by telling a single backup job to run. So at this point I am going to punt and deal w/ it later. I might try and reformat the disk. It is currently ext3 BTW, not ext4.
For anyone interested in this thread, I restarted my scheduled local backups to disk after reformatting my backup disk w/ BTRFS (starting from ext3), after which the backups have been running fine for several days. I do not know if reformatting the disk was a factor or not. I might mention that there has not been much change to the files being backed up during this time period, so not much to backup.
