LICENSE MIT
lastYesHost is already in network byte order. do not reorder it again
when displaying it.
(cherry picked from commit b5dc1903a4bca5bae099beec0b95840df7282250)
LICENSE IPL10
FIXES 20727
not our bug, but, previous code refactoring broke rx. put back keepalives in cli
ent rx connections while they are turned around.
(cherry picked from commit c34822d6105cef7f68423e4ebbf485b2d11b9635)
LICENSE IPL10
FIXES 93871
if we can't use hlist_unhashed (which is dumb) or we don't have it (less dumb) w
ork around it.
(cherry picked from commit c761f248397334400f8b4bb71cc4d01d3e4fc6ff)
LICENSE IPL10
FIXES 89557
add connection clones to allow more than maxcalls per "connection"
(cherry picked from commit f00df529b79d9df64b77dd7b054eb362a92c279e)
LICENSE IPL10
if the server is melting down or otherwise unable to reply with data, it will
cut itself off.
(cherry picked from commit c26dc0e6aaefedc55ed5c35a5744b5c01ba39ea1)
LICENSE IPL10
we learned about the peer in a previous connection... retain the information
and keep using it. widen the available window.
makes rx perform better over high latency wans. needs to be present in both
sides for maximal effect.
(cherry picked from commit 3feee9278bc8d0a22630508f3aca10835bf52866)
LICENSE IPL10
if we return leaving a non HOSTDELETED host while it does not have ALTADDR set,
a loop can happen. avoid it by simply setting ALTADDR again when we're done,
if we believe some addresses are still good this is fine.
(cherry picked from commit e6c818f070022529c362e42dfed340f54b2fd26e)
LICENSE MIT
Once we know that an attempt to stablize a buffer has failed due to
VNOVNODE, stop trying to Stablize future buffers and just invalidate
all of them belonging to the same FID.
(cherry picked from commit ef619db804d471fda3bf3f77d55e8cd4e4699b29)
LICENSE IPL10
in the event a server is "melting down" and not responding with data to RPCs
this adds the ability to time out the calls rather than letting keepalives
effectively hang the client forever. does not mark the server down since...
it's not down.
Side effect: adds similar tracking for authentication errors so tokens need
not be discarded if one server's clock has drifted.
(cherry picked from commit 0a2c144307d2471450c0660e141906a1f7b1ea36)
LICENSE IPL10
FIXES 96575
The pthread butc crashes with SIGSEGV, when I restore to a volume
which already exists. This does happen with OpenAFS 1.4.7pre4 on
i386_linux26.
Looking into the code I think butc is using the wrong function
Lp_GetRwIndex (at volser/vsprocs.c:4394). It uses the one in the
library lib/libvosadmin.a, but should use the function defined
in volser/lockprocs.c.
(cherry picked from commit 668cd5331e4857a73ba1d849446e0dfcae97eadf)
LICENSE MIT
64-bit time_t values cannot be passed to printf() for evaluation
by a %d variable. To avoid the problem we cast the value to int.
(cherry picked from commit 25d9dbb6be970d54c4711890b2085c153eda15fa)
LICENSE MIT
Add "IdleDeadTimeout" registry value and call rx_SetConnIdleDeadTime()
with it when creating new rx connections. This forces the Rx stack
to return RX_CALL_TIMEOUT if the file server has returned VBUSY for
longer than the IdleDeadTimeout.
When RX_CALL_TIMEOUT is received, the cache manager will now attempt to
find another server for the object that might be less busy.
(cherry picked from commit 2f2d5fd8a7e80802fead9ee944e33c1a0c818abd)
LICENSE IPL10
FIXES 96465
The install rule:
- installs the empty 'kdump-build' file instead of the binary on all
platforms except irix
- overwrites the selection script with the binary on linux and hpux
- doesn't have an alpha_linux branch
- installs the wrapper script as kdump32 and the binary as kdump on
solaris (instead of vice-versa)
The dest rule:
- overwrites the selection script with the binary on alpha_linux and hpux
(cherry picked from commit 7a6751c08800da5d49bf3def9175174499f6337e)
LICENSE IPL10
lightly modified from what Bren Mills submitted; the second unlock call after mfail may be entirely useless.
(cherry picked from commit b87935a4c5f26357e29f1db46eff8b1cf9185f3b)
LICENSE MIT
Implement DEBUG_REFCOUNT refcount tracking code to debug refcount issues
in the cm_buf module.
Fix a refcount leak caused by buf_IncrSyncer() incorrectly removing cm_buf_t
objects from the dirty list.
Fix the dumping of the dirty list to actually dump the dirty list.
(cherry picked from commit 088eb9158bf553d5b3dcea8f37d3bec439419d44)
LICENSE IPL10
FIXES 87977
kids these days share badly. give each one their own fd otherwise they fight.
(cherry picked from commit 6809ba29fadbd0d84ff6a9bf2ccd909b2f54b186)
LICENSE MIT
increase the cache version because the cm_volume_t data structure was modified.
(cherry picked from commit a6ef7fed731628a5e5b8af62f7e986d6f7ccf444)
LICENSE MIT
A problem was discovered with cm_Analyze when serverRef instances are
marked offline. cm_CheckOfflineVolume() was not resetting the serverRef
state. cm_Analyze was also waiting in sleep calls when it was it is
clear that waiting would not result in a change of state.
cm_CheckOfflineVolume() was updated to always reset the serverRef state
and indicate that the volume should be considered online for the purpose
of retrying when the serverRef state is reset to srv_not_busy.
The problems identified in the MIT dev.mit.edu stress test environment
are resolved by these changes.
(cherry picked from commit c3eb805fc0309efca059a26b946eb4350417867e)
