summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* By this stage of the client-read, we know a GC can't start and can't be in ↵Matthew Sackman2010-08-181-2/+1
| | | | progress. Thus even if the refcount is 0, we're still safe to read it. What's more, this area of code has never made provision for the index_lookup to return a not_found - it will crash if that happens. Thus if the implicit assumption is that the message must exist, then by the same token, we know the refcount will always be > 0
* Abstraction of the inner write functionMatthew Sackman2010-08-181-28/+26
|
* cosmeticMatthew Sackman2010-08-181-9/+8
|
* Because we can increment the refcount back above 0, the calculation of the ↵Matthew Sackman2010-08-181-1/+1
| | | | contiguous top needs to be contiguousTop + totalSize, because the validtotalsize can contain data that's well beyond the old contiguous top
* Backed out changeset c47670da4830 - I decided the dedup cache shouldn't be ↵Matthew Sackman2010-08-181-5/+7
| | | | mangled with.
* treat the dedup cache the same way as the index - permit 0 ref_counts and ↵Matthew Sackman2010-08-171-7/+5
| | | | prune on file roll
* Merging default into bug 23133Matthew Sackman2010-08-173-97/+183
|\
| * merge bug23135 into defaultMatthias Radestock2010-08-172-22/+23
| |\
| | * obtain_and_release_on_death => obtain, and minor refactor of tcp acceptorMatthew Sackman2010-08-172-16/+14
| | |
| | * Combine obtains and release_on_deathMatthew Sackman2010-08-172-30/+33
| |/
| * merge bug23132 into defaultMatthias Radestock2010-08-171-76/+165
| |\
| | * Minor correctionsMatthew Sackman2010-08-171-4/+5
| | |
| | * cosmeticMatthias Radestock2010-08-161-1/+1
| | |
| | * tweakMatthias Radestock2010-08-161-2/+2
| | |
| | * tweakMatthias Radestock2010-08-161-5/+5
| | |
| | * tweakMatthias Radestock2010-08-161-1/+1
| | |
| | * obtain_limit can be 'infinity'Matthias Radestock2010-08-161-1/+1
| | |
| | * cosmeticMatthias Radestock2010-08-161-1/+1
| | |
| | * refactor: incorporate limit checking into maybe_reduceMatthias Radestock2010-08-161-26/+22
| | |
| | * cosmeticMatthias Radestock2010-08-161-26/+35
| | |
| | * opens => open. obtains => obtainMatthew Sackman2010-08-161-61/+61
| | |
| | * Finally, it works. Important points: 1) remove elders from blocked clients ↵Matthew Sackman2010-08-161-18/+24
| | | | | | | | | | | | so that we calculate average based on pids that can actually respond. 2) reduce whenever we have pending_opens to ensure we rotate the strike.
| | * Whoops. Be consistent about comparisonsMatthew Sackman2010-08-161-1/+2
| | |
| | * Sockets have priority up to 90%. Processes have the full 100% but at lower ↵Matthew Sackman2010-08-162-55/+80
| | | | | | | | | | | | priority.
| | * Make sure tests have enough file descriptorsMatthew Sackman2010-08-161-1/+1
| | |
| | * WhoopsMatthew Sackman2010-08-161-1/+1
| | |
| | * Split obtains in two : one for open requests (higher priority) and one for ↵Matthew Sackman2010-08-161-39/+67
| | | | | | | | | | | | obtains requests (lower priority, and a lower limit)
| | * Allow the fhc server to request the client closes file handles before ↵Matthew Sackman2010-08-161-31/+52
| | | | | | | | | | | | granting an open - thus prevents deadlocks
| | * Merging default into bug 23132Matthew Sackman2010-08-162-151/+86
| | |\ | | |/ | |/|
| * | bug23129Matthias Radestock2010-08-162-151/+86
| |\ \
| | * | minor refactorMatthias Radestock2010-08-161-4/+1
| | | |
| | * | cosmeticMatthias Radestock2010-08-161-4/+3
| | | |
| | | * Obtain from the fhc sync, and obtain it before we actually open the fileMatthew Sackman2010-08-151-18/+20
| | |/ | |/|
* | | Permit 0 ref counts and don't actually forget about messages until the file ↵Matthew Sackman2010-08-161-40/+78
| |/ |/| | | | | gets removed
* | don't log file summary recover errorMatthias Radestock2010-08-151-8/+5
| | | | | | | | | | the details aren't that important and get in the way of a fresh startup
* | don't attempt to recover file summary when we know there isn't oneMatthias Radestock2010-08-151-4/+3
| |
* | cosmeticMatthias Radestock2010-08-151-4/+4
| |
* | more refactoringMatthias Radestock2010-08-151-17/+13
| |
* | yet more refactoringMatthias Radestock2010-08-151-35/+30
| |
* | now it compilesMatthias Radestock2010-08-151-12/+6
| |
* | minor refactorMatthias Radestock2010-08-151-12/+10
| |
* | cosmeticMatthias Radestock2010-08-151-1/+1
| |
* | minor refactorMatthias Radestock2010-08-151-1/+5
| |
* | Refactorings and improvents to msg_store startupMatthew Sackman2010-08-151-29/+39
| |
* | The underlying call to file:copy will not make use of any erlang-level read ↵Matthew Sackman2010-08-151-1/+1
| | | | | | | | buffer, so don't bother with it.
* | Improve, correct and document msg store recoveryMatthew Sackman2010-08-151-19/+41
| |
* | Correct and simplify the recovery from crashed compactionsMatthew Sackman2010-08-151-101/+22
| |
* | When recovering, because of the possibility of finding the same msg twice, ↵Matthew Sackman2010-08-141-4/+5
| | | | | | | | ensure that we only count and include the first copy. Other copies are then ignored.
* | A msg can appear > once in the same file because of multiple queues ↵Matthew Sackman2010-08-142-3/+5
|/ | | | receiving the same msg. When we iterate through the msgs we find in the file, we must check that the msg corresponds to the msg in the index, by matching not just with guid but also with offset. Also just for fun, assert the total_size matches. Also, fixed a mistake in file_handle_cache where the dirty flag on the destination of copy wasn't being set.
* bug 23118 merged into default (windows file descriptors limit)Marek Majkowski2010-08-131-8/+12
|\
View Lorry Controller Status