951 reputation
39
bio website
location
age
visits member for 2 years, 10 months
seen Sep 26 at 17:58

Jun
21
comment Mysql dump wont import - Created on same version of mysql etc
show us the whole line of the dump, which causes the error.
Jun
20
comment Postfix log file grows too fast
I'm glad i could help ;-)
Apr
25
comment Slower file create on faster server. Why?
You could use something like fio, but I've come across one more difference between your systems. On first systems Barracuda 7200.11 are the old, 512byte sector drives, where 7200.14 is a new Advanced format (4k sector). Metadata operations are a small writes (way smaller than 4K) and that requires the drive to read the data, modify it and write back. Not sure how xfs handles that, but i'd try passing -d su=4k,sw=1 -l su=4k to mkfs.xfs and see if that helps. btw. is your lvm on a partition? If so, please add parted -l output.
Apr
23
comment How to build a L2 Ethernet with 1000 comupters?
You usually use higher throughput core switches (10gbps if nodes use 1gbps) and usually you don't do this on L2, but rather split it into L3, that is far more manageable.
Apr
23
comment Why not block ICMP?
I believe redirects are ignored by default on linux. But yeah, better to filter that one.
Apr
23
comment Slower file create on faster server. Why?
Hmm, i'm getting out of ammunition here :( Sorry. I'd try testing it with something bit more low-level(fio), to see, if it is hardware/system problem, or if it is the filesystem. btw. both controllers in the same mode(ahci/intel)? and raid1 using md?
Apr
21
comment Slower file create on faster server. Why?
This really seems that latency gets added somewhere on the way. Are the kernels/xens the same version?
Apr
20
comment Slower file create on faster server. Why?
Can you please share mount and xfs_info on both filesystems?
Apr
20
comment Raid 0, Raid 1, Raid 5, Raid6, Raid 10 Read/Write performance?
@DukeLion thanks for the correction. Mistaken the term. Edited my answer.
Apr
20
comment Memcached in a cloud environment: dedicated server or localhosts?
How much traffic (both throughput and and reqs/s) to the memcached do you do? How important is latency to you? Are you memory-constrained on the webservers (will the data that needs to be memcached easily fit into webserver's memory)? Are you planning on any failover scenario? If so, can your app handle the data not being in cache (it should!)?
Apr
8
comment Swap 95%+ , but a lot of free ram memory
something being in swap doesen't mean its not in memory. this is called swap cache: grep SwapCached /proc/meminfo. This means that the program is in fact both, in memory and in swap. Thus when system need to free up some memory, it can just remove it instead of costly swaping it out at the moment.
Apr
8
comment SSH Kerberos Auth in Mac OSX 10.7
or you can disable strict checking and add IPs to your kerberos as well (but that's ugly)
Mar
29
comment Can Linux RAID 1 account for high latency?
FreeBSD's geom has somewhat tweakable software RAID1.
Mar
21
comment database replication for new user signup
Really depends on what you mean with high load. But since its only one single-row query and it's coming of memory, it's not that bad. You could cache the result once per second or few times in an interval equal to acceptable delay.
Mar
19
comment does the *physical* order/location of drives in a mdadm-managed RAID-10 array matter?
@locuse: interesting question ... i'm not aware of any way of backing up mdraid metadata ... you'd have to lose metadata on 2 or 3 drives to lose whole array ... actually restoring wrong metadata in wrong place would cause more harm than good. and yes, as long as those ports are in the same computer, it should work automagically :)
Mar
19
comment does the *physical* order/location of drives in a mdadm-managed RAID-10 array matter?
yes, it will reassemble correctly. that's one of the things written in the metadata ...
Jan
25
comment Reliable simple file/application distribution in a global intranet?
Well, you don't specify which protocol the app uses to check/download the updates, but if it were HTTP a well setup caching proxy could serve the purpose.
Jan
23
comment NGINX on OSX - Performing terribly slow
does setting resolver_timeout 1s; change anything? also setting error_log /Users/petrogad/logs/ngnix-sites/sandbox_log.txt debug; might bring some insight ...
Jan
23
comment ProFTPD - Failed to retrieve directory listing while using TLS
no problem. if my answer helped you, you should consider accepting it.
Jan
20
comment How to avoid the Load in rsync with ssh?
well your answer is not really mitigating the load, but rather it's impact on other apps, that's why i suggested this solution. though i'm glad you found what work's for you ;-)