MAP_DUPLICATE_KEY error in XOA backup - VM's wont START now!

Danp

@SeanMiller Have you tried running the resetvdis.py command on the affected VDIs? You can read more about it here.

jshiells

@Danp sure did, it was no help. we ended up having to restore the effected VM's from backup

there was no sign of any storage network issues during the time this started, no RT,RX errors anywhere, no traps, nothing in the zfs server logs.. nothing.... we have hundreds of VM's running , only 7 were effected and at random times during the backup process. its a little concerning..

SeanMiller

@Danp I did, I got the same results as @jshiells.

I was hoping to avoid restoring from backup as the backup is 24 hours old and changed some data between now and then..

jshiells

@SeanMiller no sorry, we ended up having to restore the VDI's form a known good backup.

SeanMiller

Im restoring from backup.

This does have me a bit concerned, This type of issue should not be caused by a backup..

olivierlambert

The root cause is not clear, I don't think it's the backup directly. It might be the generation of snapshots and reading them that cause a lot of sudden I/O on the storage.

This would require some intensive investigations, gather all the logs of all hosts and inspecting what's going on. Each installation is unique (hardware, network, disks, CPUs, memory, there's an infinite number of combinations), so finding a solution on a forum without spending days asking tons of questions seems complicated. What's concerning to me is that type of infrastructure should not be unsupported, because spending time to understand the problem is part of what we do. I'm not saying we'll find magically a solution or the root cause, but at least it could be investigated in details and compare if we have similar reports to go deeper.

jshiells

@olivierlambert I can provide information on infrastructure if it would help. the issue has not happened again though , and i cannot re create the issue in my lab. even when backups are running we are REALLY under utilizing our NFS/ZFS NVME/SSD based storage with 40gb/s links. i'm not saying its not possible, but i would be shocked. we have no recorded tx/rx errors on truenas/swtich/hosts and no record of any snmp traps coming in for any issues with all the equipment involved.

jshiells

and its happening again tonight.

I have 55 failed backups tonight with a mix of the following errors.

Error: INTERNAL_ERROR(Invalid argument: chop)
Error: SR_BACKEND_FAILURE_82(, Failed to snapshot VDI [opterr=failed to pause VDI 346298a8-cfad-4e9b-84fe-6185fd5e7fbb], )

Zero TX/RX errors on SFP's at on XEN hosts and Storage
ZERO TX/RX errors on switch ports
no errors on the ZFS/NFS storage devices
no traps in monitoring for any networking or storage issues.

olivierlambert

Well, obviously something is wrong somewhere, otherwise you wouldn't have this error. I think I never saw Error: INTERNAL_ERROR(Invalid argument: chop) in my life.

There's something triggering storage issues, not having TX/RX errors doesn't save you for other problems, like MTU problems, bad switch config or simply latency issue from the Dom0 perspective, you name it. I'd love to get it reproducible at some point but as you can see, there's "something" that makes it happen, now the question is "what". Those problems are really hard to diagnose because it's often a combo of environment and tipping points from timeouts or similar things.

tuxen

@jshiells Did you also check /var/log/kern.log for hardware errors? I'm seeing qemu process crashing with bad RIP (Instruction pointer) value which screams for a hardware issue, IMO. Just a 1-bit flipping in memory is enough to cause unpleasant surprises. I hope the servers are using ECC memory. I'd run a memtest and some CPU stress test on that server.

Some years ago, I had a two-socket Dell server with one bad core (no errors reported at boot). When the Xen scheduler ran a task on that core... Boom. Host crash.

jshiells

@olivierlambert digging more into this today, we did find this error in xensource.log related to that "CHOP" message

xensource.log.11.gz:Oct 20 10:22:08 xxx-xxx-01 xapi: [error||115 pool_db_backup_thread|Pool DB sync D:d79f115776bd|pool_db_sync] Failed to synchronise DB with host OpaqueRef:a87f2682-dd77-4a2d-aa1a-b831b1d5107f: Server_error(INTERNAL_ERROR, [ Invalid argument: chop ])

xensource.log.22.gz:Oct 19 06:06:03 fen-xcp-01 xapi: [error||27967996 :::80||backtrace] host.get_servertime D:61ad83a0cd72 failed with exception (Invalid_argument chop)

jshiells

@tuxen no sorry, great idea but we are not seeing any errors like that in kern.log. this problem when it happens is across several xen hosts all at the same time. it would be wild if all of the xen hosts were having hardware issues during the small window of time this problem happened in. if it was one xen server then i would look at hardware but its all of them, letting me believe its XOA, a BUG in xcp-ng or a storage problem (even though we have seen no errors or monitoring blips at all on the truenas server)

tuxen

@jshiells said in MAP_DUPLICATE_KEY error in XOA backup - VM's wont START now!:

@tuxen no sorry, great idea but we are not seeing any errors like that in kern.log. this problem when it happens is across several xen hosts all at the same time. it would be wild if all of the xen hosts were having hardware issues during the small window of time this problem happened in. if it was one xen server then i would look at hardware but its all of them, letting me believe its XOA, a BUG in xcp-ng or a storage problem (even though we have seen no errors or monitoring blips at all on the truenas server)

In a cluster with shared resources, It only takes one unstable host or a crashed PID left with an open shared resource to cause some unclean state cluster-wide. If the hosts and VMs aren't protected by a HA/STONITH mechanism to automatically release the resources, a qemu crash might keep one or more VDIs in an invalid, blocked state and affect SR scans done by the master. Failed SR scans may prevent SR-dependent actions (eg. VM boot, snapshots, GC kicking etc).

I'm not saying there isn't a bug somewhere but running MEMTEST/CPU tests on the host that triggered the bad RIP error would be my starting point of investigation. Just to be sure.

jshiells

@olivierlambert

After much digging I traced the following formal exceptions below from
XEN. For most of them, the "chop" error is burried in the bowels of XAPI.

ie

                Oct 18 21:46:29 fen-xcp-01 SMGC: [28249]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 238, in _parse_result
         Oct 18 21:46:29 fen-xcp-01 SMGC: [28249]     raise
Failure(result['ErrorDescription'])

It's worth noting due to the time involved in this, the only things
considered in play where nfs-01/02 and lesser degree 03. The one XCP
pool (fen-xcp), and even lesser degree XOA itself.'

I could not find any obvious issues with storage, neither hardware nor
data. Scrubs were fine, no cpu/hardware errors.

I could not find any obvious issues with xcp hosts, neither hardware nor
data. No cpu/hardware errors.

The only real change made was to correct the clock on nfs-01. I don't
see how that could affect this since most if not all locking is done
with flock.

There is a valid argument to be made that Xen technically was responding
to an issue, though not entirely clear what nor how. Most of the other
errors / wtfbbq states are either directly related (in call path), or
indirectly (xen wanted a thing, didn't get it). Those are some deep
rabbit holes.

There is more pre/post context to these, tried to include what I thought
made them a bit more easier to understand.

./SMlog.4.gz:Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]          *  E X C E
P T I O N  *
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729] leaf-coalesce: EXCEPTION <class
'XenAPI.Failure'>, ['INTERNAL_ERROR', 'Invalid argument: chop']
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 1774, in coalesceLeaf
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     self._coalesceLeaf(vdi)
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2048, in _coalesceLeaf
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     if not
self._snapshotCoalesce(vdi):
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2153, in _snapshotCoalesce
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     self._coalesce(tempSnap)
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 1962, in _coalesce
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     self.deleteVDI(vdi)
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2469, in deleteVDI
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     self._checkSlaves(vdi)
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2482, in _checkSlaves
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     self._checkSlave(hostRef, vdi)
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2491, in _checkSlave
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     text  =
_host.call_plugin(*call)
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 264, in __call__
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     return
self.__send(self.__name, args)
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 160, in xenapi_request
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     result =
_parse_result(getattr(self, methodname)(*full_params))
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 238, in _parse_result
        Oct 19 00:19:36 fen-xcp-01 SMGC: [3729]     raise
Failure(result['ErrorDescription'])

./SMlog.4.gz:Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]          *  E X C E
P T I O N  *

Oct 19 00:22:04 fen-xcp-01 SMGC: [3729] leaf-coalesce: EXCEPTION <class
'XenAPI.Failure'>, ['INTERNAL_ERROR', 'Invalid argument: chop']
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 1774, in coalesceLeaf
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     self._coalesceLeaf(vdi)
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2048, in _coalesceLeaf
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     if not
self._snapshotCoalesce(vdi):
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2153, in _snapshotCoalesce
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     self._coalesce(tempSnap)
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 1962, in _coalesce
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     self.deleteVDI(vdi)
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2469, in deleteVDI
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     self._checkSlaves(vdi)
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2482, in _checkSlaves
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     self._checkSlave(hostRef, vdi)
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2491, in _checkSlave
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     text  = _host.call_plugin(*call)
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 264, in __call__
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     return
self.__send(self.__name, args)
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 160, in xenapi_request
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     result =
_parse_result(getattr(self, methodname)(*full_params))
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 238, in _parse_result
Oct 19 00:22:04 fen-xcp-01 SMGC: [3729]     raise
Failure(result['ErrorDescription'])

./SMlog.4.gz:Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]          *  E X C E
P T I O N  *

        Oct 19 00:22:11 fen-xcp-01 SMGC: [3729] gc: EXCEPTION <class
'XenAPI.Failure'>, ['INTERNAL_ERROR', 'Invalid argument: chop']
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 3388, in gc
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]     _gc(None, srUuid,
dryRun)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 3273, in _gc
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]     _gcLoop(sr, dryRun)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 3214, in _gcLoop
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]
sr.garbageCollect(dryRun)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 1794, in garbageCollect
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]
self.deleteVDIs(vdiList)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2374, in deleteVDIs
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]     SR.deleteVDIs(self,
vdiList)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 1808, in deleteVDIs
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]     self.deleteVDI(vdi)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2469, in deleteVDI
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]     self._checkSlaves(vdi)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2482, in _checkSlaves
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]
self._checkSlave(hostRef, vdi)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/opt/xensource/sm/cleanup.py", line 2491, in _checkSlave
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]     text  =
_host.call_plugin(*call)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 264, in __call__
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]     return
self.__send(self.__name, args)
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 160, in xenapi_request
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]     result =
_parse_result(getattr(self, methodname)(*full_params))
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 238, in _parse_result
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]     raise
Failure(result['ErrorDescription'])
         Oct 19 00:22:11 fen-xcp-01 SMGC: [3729]



./SMlog.5.gz:Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]          *  E X C
E P T I O N  *
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714] GC process exiting, no work left
        Oct 18 21:51:23 fen-xcp-01 SM: [30714] lock: released
/var/lock/sm/0cff5362-5c89-2241-2207-a1d736d9ef5e/gc_active
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714] In cleanup
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714] SR 0cff ('fen-nfs-03 - DR
(Diaster Recovery Storage ZFS/NFS)') (608 VDIs in 524 VHD trees): no changes
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]
*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]          ***********************
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]          *  E X C E P T I O N  *
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]          ***********************
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714] gc: EXCEPTION <class
'XenAPI.Failure'>, ['INTERNAL_ERROR', 'Invalid argument: chop']
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/opt/xensource/sm/cleanup.py", line 3388, in gc
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     _gc(None, srUuid, dryRun)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/opt/xensource/sm/cleanup.py", line 3273, in _gc
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     _gcLoop(sr, dryRun)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/opt/xensource/sm/cleanup.py", line 3214, in _gcLoop
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     sr.garbageCollect(dryRun)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/opt/xensource/sm/cleanup.py", line 1794, in garbageCollect
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     self.deleteVDIs(vdiList)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/opt/xensource/sm/cleanup.py", line 2374, in deleteVDIs
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     SR.deleteVDIs(self, vdiList)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/opt/xensource/sm/cleanup.py", line 1808, in deleteVDIs
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     self.deleteVDI(vdi)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/opt/xensource/sm/cleanup.py", line 2469, in deleteVDI
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     self._checkSlaves(vdi)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/opt/xensource/sm/cleanup.py", line 2482, in _checkSlaves
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     self._checkSlave(hostRef, vdi)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/opt/xensource/sm/cleanup.py", line 2491, in _checkSlave
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     text  =
_host.call_plugin(*call)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 264, in __call__
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     return
self.__send(self.__name, args)
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 160, in xenapi_request
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     result =
_parse_result(getattr(self, methodname)(*full_params))
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]   File
"/usr/lib/python2.7/site-packages/XenAPI.py", line 238, in _parse_result
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]     raise
Failure(result['ErrorDescription'])
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]
*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714] * * * * * SR
0cff5362-5c89-2241-2207-a1d736d9ef5e: ERROR
        Oct 18 21:51:23 fen-xcp-01 SMGC: [30714]
        Oct 18 21:51:28 fen-xcp-01 SM: [26746] lock: opening lock file
/var/lock/sm/894e5d0d-c100-be00-4fc4-b0c6db478a26/sr

olivierlambert

Well, changing time on the NFS server could have an impact, since file access time attributes are used to do many checks in general. I'm not saying that's 100% the issue, but it tells how deep a small change could cause havoc.

jshiells

@olivierlambert

Just an update on this:

we made sure all our server times were synced. issue happened again the next run.

just for shits and giggles we restarted toolstack on the all the hosts yesterday and the problem went away. no issues with the backup last night. maybe just a coincidence, we are continuing to monitor.

we also noticed that even though this CHOP error is coming up, snapshots are getting created

jshiells

I have, what is hopefully a final update to this issue(s).

we upgraded to xoa version .99 a few weeks ago and the problem has now gone away. we suspect that some changes were made for timeouts in xoa that have resolved this , and a few other related problems.