Help Needed: Long Spool Enumeration Time on Service Restart

What is your average message size?

I’m guessing 130k.

In that IOstat output your idle was low, which means your drive is running relatively close to capacity.

yes around this

So in the image you’re showing around 12,000 kb/s, that seems rather slow for a disk.

okay so problem is related to my diska

disk

so what can i do here?

This isn’t an issue I’m otherwise hearing about, so I’d start by securing yourself a faster spool disk. You’re in iowait 82% of the time. Your CPU is waiting around for the disk.

so i have to replace the disk with the better one?

or is there anyother way?

Unless you can make that existing disk faster somehow.

You may also want to try wiping the spool entirely and starting again in case you have a corrupt rocksdb.

As in shut down, delete the spool directories, recreate them, start.

im not using rocksdb im using local disk for spooling

Well I wouldn’t know that since you ignored the READ ME FIRST and didn’t post your config.

if i delete the spool file then it will also delte the mails. i mean queued mails

am i right or wrong here/

?

Yes, or you can wait for them to spool in.