Error - No space left using (DRBD, POSTFIX, DOVECOT, OCFS2, O2CB) -dovecot-lmtp

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Error - No space left using (DRBD, POSTFIX, DOVECOT, OCFS2, O2CB) -dovecot-lmtp

noeldiazqwerty
Good evening Sirs:

I have a drbd 8.4 running in a email-production server, but lately I have found that it has just 64% of use at home-partition, and 64% of i-nodes in that partition too, but meanwhile dovecot and postfix are running writing data on each nodes on a primary-primary roles Dovecot says at my dovecot-lmtp:

"No space left on device", I want to know why if I have 1 terabyte around and below 100% of use as I mentioned upstairs lines, when this occurs my email production server collapses at all, and all users notifies that they can not see their data, tells me that their messages were sent but server couldnot make a copy or save at draft folders even sent folder, I have used many solutions by forums, I have talked to Mr Witese Venema on this course of problem, but rather I havenot given the apropietaly solution. I have used to stand by the victim node and put online the other one, and visceverse after data-synchronization process.

I am afraid that if i develop a new email-server with drbd  10.0 to debian 9 or 10,  what do I do if it  happens the same?, that my server has 1 Tb and backs to tell me the same not even used whole data as case before?.

Thanks you, any answer, or brightly idea will be taken as a big solution from me.
Reply | Threaded
Open this post in threaded view
|

Re: Error - No space left using (DRBD, POSTFIX, DOVECOT, OCFS2, O2CB) -dovecot-lmtp

Wietse Venema
Noel Diaz:

> Good evening Sirs:
>
> I have a drbd 8.4 running in a email-production server, but lately I have
> found that it has just 64% of use at home-partition, and 64% of i-nodes in
> that partition too, but meanwhile dovecot and postfix are running writing
> data on each nodes on a primary-primary roles Dovecot says at my
> dovecot-lmtp:
>
> "No space left on device", I want to know why if I have 1 terabyte around
> and below 100% of use as I mentioned upstairs lines, when this occurs my
> email production server collapses at all, and all users notifies that they
> can not see their data, tells me that their messages were sent but server
> couldnot make a copy or save at draft folders even sent folder, I have used
> many solutions by forums, I have talked to Mr Witese Venema on this course
> of problem, but rather I havenot given the apropietaly solution. I have
> used to stand by the victim node and put online the other one, and
> visceverse after data-synchronization process.

Dovecot says "no space left on device", because the file system
reports that error to Dovecot. Dovecot is just the messenger of vad
news, don't shoot the messenger.

So the problem is in your file system (ocfs or whatever, on top of
DRBD). I did a quick websearch, and found many web pages that mention
this exact error.

> I am afraid that if i develop a new email-server with drbd  10.0 to debian
> 9 or 10,  what do I do if it  happens the same?, that my server has 1 Tb
> and backs to tell me the same not even used whole data as case before?.

My opinion is that either the file system technology is not ready
for what you want to do, or you are making some fundamental mistake.

If it is a technology problem, consider sharing the file system
using an NFS server on top of a RAID array.

If you are using dual-primary DRBD outside a LAN environment, or
using ext4 instead of something like ocfs2, then you are making a
fundamental mistake. You WILL have file system corruption.

> Thanks you, any answer, or brightly idea will be taken as a big solution
> from me.

Since the error happens far outside of Postfix and Dovecot, it is hard
to give a solution other than to stop doing what you're doing.

        Wietse