Restoring reseller level backup of useraccount takes forever

[Invader Zim]

Ever since the last update we've had issues restoring user accounts. Since we're currently moving users from an old server to a new server this is kind of annoying. Wether we restore the backup of 1 user weighing in at 10Mb or 10 users at a total of 4Gb it takes about 20 minutes either way for the restore to complete.

During this time the CPU consumption of dataskq runs up to 70% for about 15 - 18 minutes while seemingly nothing else happens.

 

[DirectAdmin Support]

Hello,

To check to see what it's up to, type:

ps ax
killall -USR1 dataskq
tail /var/log/directadmin/errortaskq.log

usually the last few entrie of the ps output would be any binaries that the dataskq may have called (usually tar for extraction). The killall -USr1 dataskq and tail will get the dataskq to dump it's current location to the /var/log/directadmin/errortaskq.log, where we could then determine what area it's chewing on to see if there is any particular reason it's gonig slower than usual (eg: email us the pasted output)

John

 

[Invader Zim]

This evening we're going to be moving the last users out to the new server. I'll do that when we're done.

As for now, the only real error I noticed in the errortaskq.log is

ConfigFile::writeFile(`) : Timestamp from when it was read is different, overwriting anyway

 

[DirectAdmin Support]

that's more of a warning saying that the dataskq has higher importance than anything else.. including browsing in DA and will ovewrite the updated values regardless of what the directadmin binary is doing.

John