Hi Karthik,
Thank you for the response. I've pass this information to the customer. They replied and have another inquiry:
-----------------------------
Hello,
Based on “It is the genric workflow. In case of call_bail after frame-timeout, client will send a disconnect and try to reconnect.” feedback:
I made a test in glusterfs and created a BLOCKED posix lock on a file, which was hold more than “frame-timeout” seconds. But it did not create a gluster client disconnection. Could you explain the “call bail” mechanism and either BLOCKED inode locks or posix locks trigger it?
DUMP-START-TIME: 2022-12-05 18:29:24.538081
path=/vTeMIPA_director/vTeMIP_data/locks/test.log
mandatory=0
posixlk-count=2
posixlk.posixlk0=type=WRITE, whence=0, start=0, len=0, pid = 4659, owner=800f867c941bdbff, client=0x7faa84007460, connection-id=CTX_ID:49110287-f46f-4e3c-b8ea-70c1bd0c200a-GRAPH_ID:0-PID:23066-HOST:osst127.gre.hpecorp.net-PC_NAME:vTeMIPA_dirdata-client-0-RECON_NO:-3, blocked at 2022-12-05 18:18:23, granted at 2022-12-05 18:18:23
posixlk.posixlk1=type=WRITE, whence=0, start=0, len=0, pid = 5453, owner=4bbdfd655f08860e, client=0x7faa84007460, connection-id=CTX_ID:49110287-f46f-4e3c-b8ea-70c1bd0c200a-GRAPH_ID:0-PID:23066-HOST:osst127.gre.hpecorp.net-PC_NAME:vTeMIPA_dirdata-client-0-RECON_NO:-3, blocked at 2022-12-05 18:19:59
[root@osss211 ~]# gluster v get vTeMIPB_dirdata all | grep -i frame
network.frame-timeout 180
No “call_bail” error, no “Transport Endpoint is not connected” error in gluster client log.
--------------------------------