• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

/tmp for migration unpack??

I

IMspintheweb2

Guest
/tmp isnt big enough for most of the domains I want to migrate...

Where's the script that creates the xml that saysto unpack in /tmp?

Can't that be changed easily to use /var/tmp ?

tia!
 
Not sure that it is possible because it is hardcoded in PMM. I think that you can just try to create mountpoint for /tmp on other big partition.
 
that's too bad - I wasn't looking forward to re partitioning /tmp
 
Prior to 9.2.x this was not how it worked, whoever hard coded it has no idea what they're doing but it works great at causing migrations to fail and then of course when /tmp is full the control panel doesn't respond at all, requests just hang, so you have to go clean it up manually. The same person is probably the one that didn't bother to check how much space was in /tmp before extracting to it.
 
I have submitted request to developers regarding possibility for changing default /tmp path. I will update this thread as soon as I receive any useful information.
 
tmp out of disk space

I also see this problem in Plesk 9.2.3 on Centos 5.6, while attempting to do a migration. The workaround mentioned above (pmm-ras replacement) does NOT work.
 
tmp out of disk space

As indicated in the original post at the top of the thread, /tmp is used for migration unpack and plesk runs out of disk space when /tmp is not sufficiently large. in my case, i have a 2G /tmp partition and the migration data is about 15 times that size. this problem still happens after using the replacement pmm-ras, it would be solved if migration agent used a configurable path somewhere else other than /tmp.
 
I have informed developers with your information and will keep you updated with their answer in this thread.
 
Developers have informed that it is old known problem. Bug already submitted and fix will be included to the one of nearest Plesk update.
 
one simple workaround

OK, thanks. For my systems (Centos 5.4), I found an easy workaround which might be useful to other experiencing the same when performing a migration.

1. make a copy of etc/fstab, edit it and remove the mount line for /tmp
2. reboot and chmod 777 /tmp
3. restart psa and mysqld
4. perform migration and verify results
5. copy back the old /etc/fstab in place and reboot

Alternately, a bigger device can be made and mounted for tmp, but that is more complicated and unnecessary since the migration only needs to be performed once.
 
Hi,

I have a REALLY old server that I'm trying to migrate a site from (about 500MB) and I've come across this problem. My root partition is only about 400MB. I've tried your technique and it hasn't worked, my /etc/hda1 partition is still filling up and freezing the migration.

Can you give me any more pointers?
 
Ah, in case it helps anyone it looks like Plesk 7.5.4 uses a /migration so I needed to link that to a bigger partition. All sorted!
 
Back
Top