Single-Copy Message storage ?

[tiger2000]

Dear All,

Just hope to clarify this feature : Single-Copy Message storage :

if someone sends an email with 10MB size to 1000 recipients in system, does it mean (by "Single-Copy Message Storage" definition) ZCS will only consume 10MB (or probably less than 20MB) size in file system?

Is this correct?

[Klug]

Yes, as long as all the users are on the same server.

[tiger2000]

just did a test :

1. create 300 accounts and assign all of them into a distribution list.

2. send an email to such DL with size : 20MB , thus it should consume about 20MB x 300 = 6000 MB size in other email system.

3. check the filesystem df result during email delivery:

/dev/hda1 20161172 5519900 13617132 29% /
/dev/hda1 20161172 5519904 13617128 29% /
/dev/hda1 20161172 5519908 13617124 29% /
/dev/hda1 20161172 5519908 13617124 29% /
/dev/hda1 20161172 5519908 13617124 29% /
/dev/hda1 20161172 5519908 13617124 29% /
/dev/hda1 20161172 5519908 13617124 29% /
/dev/hda1 20161172 5524624 13612408 29% /
/dev/hda1 20161172 5533736 13603296 29% /
/dev/hda1 20161172 5540496 13596536 29% /
/dev/hda1 20161172 5612420 13524612 30% /
/dev/hda1 20161172 5669804 13467228 30% /
/dev/hda1 20161172 5702168 13434864 30% /
/dev/hda1 20161172 5754112 13382920 31% /
/dev/hda1 20161172 5741920 13395112 31% /
/dev/hda1 20161172 5760788 13376244 31% /
/dev/hda1 20161172 5761320 13375712 31% /
/dev/hda1 20161172 5761320 13375712 31% /
/dev/hda1 20161172 5761320 13375712 31% /
/dev/hda1 20161172 5761320 13375712 31% /
/dev/hda1 20161172 5761320 13375712 31% /
/dev/hda1 20161172 5781872 13355160 31% /
/dev/hda1 20161172 5781872 13355160 31% /
/dev/hda1 20161172 5781872 13355160 31% /
/dev/hda1 20161172 5781872 13355160 31% /
/dev/hda1 20161172 5781880 13355152 31% /
/dev/hda1 20161172 5781880 13355152 31% /
/dev/hda1 20161172 5789532 13347500 31% /
/dev/hda1 20161172 5848040 13288992 31% /
/dev/hda1 20161172 5888824 13248208 31% /
/dev/hda1 20161172 5884696 13252336 31% /
/dev/hda1 20161172 5889696 13247336 31% /
/dev/hda1 20161172 5905248 13231784 31% /
/dev/hda1 20161172 5905248 13231784 31% /
/dev/hda1 20161172 5905248 13231784 31% /
/dev/hda1 20161172 5905248 13231784 31% /
/dev/hda1 20161172 5885260 13251772 31% /
/dev/hda1 20161172 5822464 13314568 31% /
/dev/hda1 20161172 5860996 13276036 31% /
/dev/hda1 20161172 5866360 13270672 31% /
/dev/hda1 20161172 5866360 13270672 31% /
/dev/hda1 20161172 5866360 13270672 31% /

looks like : it's consuming <300MB size in file system, although it's already much smaller than 6GB , it doesn't act as what the document says.

is my result normal behavior ? or any advice ?

Thanks.

[Rich Graves]

du -skh index store redolog db

store (the actual message) is deduped. You get one 24MB copy (with base64 overhead) there.

redolog saves a copy for crash recovery... then either throws it away upon log switch on OSS edition, or archives it for NE point-in-time restores.

index is not deduped, afaik. Each of those 300 new users would initialize their index. Even a mostly empty index has a minimum size. Send a second 20MB attachment to the same 300 users, and I'd expect a smaller increase, closer to 50MB (MIME overhead+redolog).

I wouldn't expect db/data to grow appreciably, but the first message to a mailbox(group) might trigger something.