Commit Graph

382 Commits

Author SHA1 Message Date
Alexander Motin
4ce7a0868c Remove duplicate and incomplete code handling LUN modify.
Instead reuse code from LUN creation.  This allows most of LUN media
options to be changed live with modify request without full restart.
2015-09-22 10:45:50 +00:00
Alexander Motin
8153ec9ad5 Log iSCSI session reinstatements.
False session reinstatements can be result of misconfiguration, when
several initiators use the same initiator name and ISID.
2015-09-21 13:25:29 +00:00
Alexander Motin
218d25d00b Mark with DMA flag I/Os waiting for iSCSI write data after R2T.
Reads and immediate writes are not blocking, so don't bother.
2015-09-20 11:39:19 +00:00
Alexander Motin
b22213694e Remove couple excess SGLIST I/O flags.
Those flags duplicated respective (sg_entries > 0) values.
2015-09-20 10:40:30 +00:00
Alexander Motin
921af50782 Always execute REPORT LUNS and REQUEST SENSE commands locally.
REPORT LUNS command is more related to target rather then specific LUN.
This node may be primary for LUNs for some reason unknown to another,
and command forwarded to another node won't be able to report them.

REQUEST SENSE is related to LUN, but in our implementation it reports
only UAs and CAs, that are stored locally rather then on primary node.
2015-09-19 19:30:55 +00:00
Alexander Motin
efbf6139a4 Split two command flags with different meaning.
This is only a cosmetical change.
2015-09-19 19:11:59 +00:00
Alexander Motin
027dd0cfd0 Pack struct ctl_ha_msg_hdr by 8 bytes. 2015-09-19 18:23:06 +00:00
Alexander Motin
6861045bda Fix memory corruption when >128K transferred through HA link.
While there, do some cleaning.
2015-09-19 13:12:09 +00:00
Alexander Motin
bc8749e2ff Mark I/Os with DMA flag while moving data through the HA link. 2015-09-18 20:11:10 +00:00
Alexander Motin
75a3108e13 Relax serseq option operation for reads.
Previously, with serseq enabled, next command was unblocked only after
previous completed.  With this change, for read operations, next command
is unblocked as soon as last media read completed.  This is important
for frontends that actually wait for data move completion (like camtgt),
or when data are moved through the HA link, or especially when both.
2015-09-18 19:43:14 +00:00
Alexander Motin
59bb97a934 Kill HA link and shutdown the threads on shutdown. 2015-09-18 12:08:00 +00:00
Alexander Motin
0c05f0dc16 Replicate initiators WWPNs and names between HA peers. 2015-09-17 21:51:11 +00:00
Alexander Motin
54713bce25 Replicate port->init_devid to HA peer. 2015-09-17 20:36:34 +00:00
Alexander Motin
6213882769 When reporting TPT UA, report which of thresholds was reached. 2015-09-17 17:00:36 +00:00
Alexander Motin
042e9bdc41 Report number of failed XCOPY segment. 2015-09-17 14:22:52 +00:00
Alexander Motin
7f7bb97a0f Report proper medium error code for VERIFY commands. 2015-09-17 12:52:18 +00:00
Alexander Motin
83981e319d Fix reading after end of file for file-backed LUNs.
If backing file is smaller then the LUN size, we have to explicitly clear
the rest of the buffer to not leak some random data from previous I/Os.
2015-09-16 21:43:51 +00:00
Alexander Motin
d6043e4643 Make COMPARE AND WRITE report offset of difference. 2015-09-16 18:33:04 +00:00
Alexander Motin
5c5fb901eb Frontends don't need to set errors themselves. 2015-09-16 13:25:35 +00:00
Alexander Motin
e2c3044b13 Don't flap the HA link if sysctl is reset to the same value. 2015-09-16 09:59:05 +00:00
Alexander Motin
18e87f95f7 Fix completion/error status reporting. 2015-09-15 10:42:53 +00:00
Alexander Motin
923863546a Close potential race between datamove and HA failover. 2015-09-15 09:36:46 +00:00
Alexander Motin
d0f25804a7 Implement data/status aggregation for secondary HA node.
For short read requests this reduces latency by 30%, reporting command
completion after two interlink roundtrips instead of full three.
2015-09-14 15:21:23 +00:00
Alexander Motin
3ed019a919 Implement data/status aggregation for camsim backend.
This is almost pointless for primary HA node, but useful for secondary,
where between fe_datamove and fe_done calls goes another link roundtrip.
2015-09-14 15:15:32 +00:00
Alexander Motin
7b6660c9aa Remove CTL_PRIV_LBA_LEN from HA messages.
Previously it was used for statistics, but now just a 16 extra bytes.
2015-09-14 09:56:01 +00:00
Alexander Motin
ceff31dc0c Implement QUERY TASK, QUERY TASK SET and QUERY ASYNC EVENT.
Now we support most of SAM-5 task management.
2015-09-14 08:01:05 +00:00
Alexander Motin
bb978da493 Map CLEAR TASK SET and I_T NEXUS RESET for iSCSI.
The last should not be called without iSCSIProtocolLevel negotiation.
2015-09-13 20:40:00 +00:00
Alexander Motin
4cbc0bc5ae Implement iSCSI TARGET COLD RESET task management function.
Implement it as CTL_TASK_TARGET_RESET plus termination of all sessions.
2015-09-13 20:22:18 +00:00
Alexander Motin
6187d4722a Improve read-only support. 2015-09-13 16:49:41 +00:00
Alexander Motin
6b56dd0cfa Reannounce port to HA peer if LUN map changed after online. 2015-09-13 15:31:55 +00:00
Alexander Motin
cfa0987386 Report INQUIRY DATA HAS CHANGED for related LUNs on port on-/offline. 2015-09-13 14:52:31 +00:00
Alexander Motin
eebde34629 Report CTL_UA_LUN_CHANGE on LUN map change. 2015-09-13 14:04:42 +00:00
Alexander Motin
844c2a3a94 Make TARGET RESET respect LUN mapping. 2015-09-13 13:00:20 +00:00
Alexander Motin
84e2fad15a Check for obsolete NUL bin in CSCD descriptor. 2015-09-12 20:45:09 +00:00
Alexander Motin
051481a24e Add HA support for CTL_TASK_I_T_NEXUS_RESET. 2015-09-12 20:06:22 +00:00
Alexander Motin
0a4c638188 Fix false CTL_UA_RES_RELEASE on secondary HA node. 2015-09-12 18:29:05 +00:00
Alexander Motin
a65a997fd9 Improve XCOPY error reporting. 2015-09-12 16:30:01 +00:00
Alexander Motin
238b6b7c75 Report that we have no limit on POPULATE TOKEN segment size. 2015-09-12 14:20:11 +00:00
Alexander Motin
eed0d596cf Correct RTPG bitmask. 2015-09-12 13:53:41 +00:00
Alexander Motin
e184d2576a Some HA polishing. 2015-09-12 12:46:04 +00:00
Alexander Motin
d36c617616 CTL documentation update, mostly for HA. 2015-09-12 10:23:23 +00:00
Alexander Motin
572846e089 Make ctld restart on primary HA node less invasive for secondary. 2015-09-11 14:47:35 +00:00
Alexander Motin
ee4ad294d2 Close races between device close and request processing.
All requests arriving for processing after OFFLINE flag set are rejected
with BUSY status.  Races around OFFLINE flag setting are closed by calling
taskqueue_drain_all().
2015-09-11 14:33:05 +00:00
Alexander Motin
3236151ea8 Reference/release devices on every I/O, rather on open/close.
While this may be slower, it allows device destruction to complete,
rather then block waiting for indefinitely long time.
2015-09-11 12:50:52 +00:00
Alexander Motin
7ac58230ea Reimplement CTL High Availability.
CTL HA functionality was originally implemented by Copan many years ago,
but large part of the sources was never published.  This change includes
clean room implementation of the missing code and fixes for many bugs.

This code supports dual-node HA with ALUA in four modes:
 - Active/Unavailable without interlink between nodes;
 - Active/Standby with second node handling only basic LUN discovery and
reservation, synchronizing with the first node through the interlink;
 - Active/Active with both nodes processing commands and accessing the
backing storage, synchronizing with the first node through the interlink;
 - Active/Active with second node working as proxy, transfering all
commands to the first node for execution through the interlink.

Unlike original Copan's implementation, depending on specific hardware,
this code uses simple custom TCP-based protocol for interlink.  It has
no authentication, so it should never be enabled on public interfaces.

The code may still need some polishing, but generally it is functional.

Relnotes:	yes
Sponsored by:	iXsystems, Inc.
2015-09-10 12:40:31 +00:00
Alexander Motin
fb606ebabc Remove unused target and initiator IDs. 2015-09-10 10:46:21 +00:00
Alexander Motin
3f326305fc Disable CTL_IO_DELAY feature.
It is too developer-oriented to be enabled by default.
2015-09-10 09:27:22 +00:00
Alexander Motin
a3977bea20 Allow LUN options modification via CTL_LUNREQ_MODIFY.
Not all changes take effect, but that is a different question.
2015-09-06 11:23:01 +00:00
Alexander Motin
0bcd4ab6ba Move setting of media parameters inside open routines.
This is preparation for possibility to open/close media several times
per LUN life cycle.  While there, rename variables to reduce confusion.
As additional bonus this allows to open read-only media, such as ZFS
snapshots.
2015-09-06 09:54:56 +00:00
Alexander Motin
bd236ba5c0 Remove some dead code. 2015-09-04 09:19:01 +00:00