Hi, On Sun, 29 Aug 2021, at 8:28 PM, Lists Nethead wrote: > 29392: write(6,"S2 SYNCAPPLY MAILBOX %(UNIQUEID > 6uxzsiuvofrhrf69dfs1uqxq MBOXNAME user.ppgw SYNC_CRC 2538953411 > SYNC_CRC_ANNOT 307862686 LAST_UI"...,4096) = 4096 (0x1000) > 29392: write(6,"ATIONS (%(ENTRY /vendor/cmu/cyrus-imapd/thrid USERID > NIL VALUE 87f494ec1d410ff7))) %(UID 18 MODSEQ 194 LAST_UPDATED > 1620897300 F"...,4096) = 4096 (0x1000) > 29392: write(6,"12 LAST_UPDATED 1620897300 FLAGS () INTERNALDATE > 1608458977 SIZE 58108 GUID 03fd1fb95a2083a702fcf679a0e62182698ea1f8 > ANNOTATIONS"...,1984) = 1984 (0x7c0) It's hard to see what it might be complaining about, since the "S2 SYNCAPPLY ..." command it's trying to send is 4096+4096+1984 bytes long, and is heavily abbreviated in the strace output. You can log what's actually going over the wire by enabling telemetry for the user: https://www.cyrusimap.org/imap/reference/faqs/o-telemetry.html In this case, you would want to set up telemetry on your replica server, for the user named by "sync_authname" on your main server. This will make the replica log all the replication traffic it receives. If you can reproduce the bad protocol error by replicating just one test folder for one test user, then that's probably the least-noisy way to proceed, and without worry about exposing anyone's mail contents. But if you can't reproduce it with a test folder in a test account, then it might be something specific to the user or mailbox that's causing the problem, which will be harder to figure out... Anyway, the first step is capturing some telemetry. Cheers, ellie ------------------------------------------ Cyrus: Info Permalink: https://cyrus.topicbox.com/groups/info/Te28e46bb80b596a6-M515bb1976a2a5e4fa7ed418d Delivery options: https://cyrus.topicbox.com/groups/info/subscription
